How Do Equity Markets Price Disaster Risk? Evidence from Vale after the Mariana Disaster

How Do Equity Markets Price Disaster Risk? Evidence from Vale after the Mariana Disaster | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article How Do Equity Markets Price Disaster Risk? Evidence from Vale after the Mariana Disaster Matheus Soares Mendes, João Antônio da Costa Neto, Francisco de Assis Miranda da Silva, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8894226/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 9 You are reading this latest preprint version Abstract This study examines how the Fundão tailings dam collapse (Mariana, Brazil) was priced in Vale S.A.’s equity returns, and whether the observed dynamics are consistent with rapid adjustment under the Efficient Market Hypothesis (EMH) when the shock has ‘black swan’ features. We apply the synthetic control method to construct a transparent counterfactual path for Vale’s semiannual cumulative stock returns over 2011.2-2017.2. To address potential spillovers and comparator contamination, we estimate three donor-pool specifications: DP1 (Brazilian and international peers), DP2 (excluding Brazilian firms), and DP3 (excluding Brazilian firms and BHP Group). DP1 delivers the strongest pre-disaster fit and is adopted as the baseline. The results show a negative return response in the disaster semester (2015.2), consistent with an immediate shock, but the most informative pattern arises afterward: Vale underperforms its synthetic counterfactual persistently in 2016.1-2017.2. This sustained post-event gap suggests delayed market adjustment as legal, regulatory, and reputational uncertainties unfold beyond the initial news window. Placebo-style comparisons support the interpretation that Vale’s post-disaster divergence is not a mere artifact of random fluctuations under strong pre-event fit. Overall, the study contributes to disaster finance by providing firm-level evidence on the persistence of disaster-related repricing and by highlighting governance, transparency, and disaster-risk management as relevant mechanisms for limiting valuation losses under extreme events. JEL classification: G14; Q54; Q56; L72; C23 Mariana disaster tailings dam collapse synthetic control stock returns disaster finance market efficiency Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. Introduction A decade after the collapse of the Fundão dam, the Mariana disaster remains a critical milestone in Brazil’s mining history, given the persistent legal and social challenges. Although a landmark R $ 170 billion (≈ US $ 27.5 billion) reparation agreement, signed between mining companies and public authorities, was finalized in late 2024, legal battles continue to unfold both domestically and internationally. 1 By the end of 2025, no criminal convictions had been reached in Brazil; meanwhile, in the United Kingdom, a civil action against BHP (listed on the London Stock Exchange at the time) had its first phase concluded in March 2025, but remains pending final judgment (Mansur, 2025 ). These enduring repercussions show that, although such events are episodic, their environmental, financial, and legal effects persist over the long term, directly affecting the governance and valuation of the sector’s main players. In the disaster-economics literature, this persistence is expected: disasters generate direct and indirect economic losses and can raise perceived risk and risk premia, with consequences that extend beyond the immediate physical shock (Maran, 2023 ). The relevance of this impact is underscored by the strategic position that mining occupies within the Brazilian economy, standing out as one of the most significant sectors in both financial markets and employment generation. In 2024, the sector represented about 4% of Brazil’s Gross Domestic Product (GDP), handling approximately 1.24 billion tons of minerals and creating over 210,000 direct jobs and an additional 2.2 million indirect and induced jobs (IBRAM, 2024). Additionally, the sector was responsible for collecting R $ 85.6 billion (≈ US $ 13.8 billion) in taxes and R $ 6.86 billion (≈ US $ 1.1 billion) in Financial Compensation for Mineral Exploitation (CFEM), highlighting its economic significance. The leading minerals produced include iron ore, gold, and copper (IBRAM, 2024). Against this backdrop, a natural question arises: how do extreme events (environmental disasters) affect the stock market valuation of firms operating in such a strategic sector? Within this setting, Vale S.A., founded in 1942, is the primary representative of the mining sector in Brazil and one of the largest global mining companies. It currently holds a world-leading position in iron ore production, with shares listed on the São Paulo, Paris, Madrid, and New York stock exchanges (Fabrício et al., 2021 ). However, this economic prominence comes with considerable operational challenges, especially concerning tailings dam safety. Globally, between 1915 and 2019, there were 356 dam failures recorded, reflecting the gravity and frequency of such operational risks (WMTF, 2019). These incidents typically stem from severe natural events, such as earthquakes and storms, or structural and planning failures, including operational and maintenance shortcomings (Duarte, 2008 ). Beyond the tragic human and environmental toll, the finance-oriented disaster literature stresses that shocks of this nature may also trigger liquidity pressure and heightened risk assessments, with spillovers to corporate financial conditions and market pricing (Halkos & Zisiadou, 2021 ). In Brazil, particularly in the state of Minas Gerais, the risks associated with tailings dams are frequent and severe. Between 1986 and 2019, at least seven major dam failures were recorded in the state. Among these, the Fundão dam rupture in 2015 (controlled by Samarco, a joint venture between Vale and BHP Billiton, in Mariana) stands out as the most catastrophic. This event was Brazil’s worst environmental disaster, causing 19 fatalities, widespread community destruction, and severe river contamination (Fernandes et al., 2016 ). Such events, often framed as “black swans,” are rare, difficult to anticipate, and can generate extraordinary impacts that test standard assumptions about risk and information processing in financial markets (Taleb, 2007 ; Woods, 2006 ; Masys, 2012 ). Brazil, recognized globally as a major commodities exporter with prominent natural resource-based companies, has repeatedly faced serious environmental incidents that affected corporate valuations. Notably, Vale itself experienced another devastating dam collapse in Brumadinho in 2019. Moreover, the oil spills in the Campos Basin involving Chevron in 2011 and Petrobras in 2019 further illustrate how environmental disasters can influence corporate stock prices. In disaster economics, a recurring point is that recovery and post-shock adjustment depend critically on financial channels (access to finance, borrowing, and the broader functioning of credit markets) which helps explain why market-based outcomes may react sharply and, at times, persistently (Keerthiratne & Tol, 2017 ). In parallel, the literature also discusses risk-mitigation and resilience instruments (e.g., insurance and related financial tools) as part of the policy and governance toolkit (Halkos & Zisiadou, 2021 ). Considering these extreme events, a critical question follows: to what extent can financial markets rapidly price information related to unexpected crises? The Efficient Market Hypothesis (EMH) posits that markets quickly absorb all available information and immediately reflect it in asset prices (Fama, 1970 ; 1991 ). However, rare and high-impact events inherently challenge this view, raising doubts about the market’s ability to anticipate and price such occurrences accurately in real time (Taleb, 2007 ; Woods, 2006 ; Masys, 2012 ). Several prior studies have investigated the socioeconomic and environmental impacts of these disasters, including works by Simonato ( 2017 ), Castro and Almeida ( 2019 ; 2023 ), Carrillo et al. ( 2020 ), and Biazoli et al. ( 2024 ). Recently, Biazoli et al. ( 2024 ) advanced this field by applying cluster and synthetic control methods to assess the regional economic impacts of the Mariana disaster. Nevertheless, studies have not specifically examined the isolated effects of the Mariana disaster on Vale S.A.’s stock returns, leaving a relevant gap at the intersection of disaster economics and capital markets. This study seeks to fill that gap by evaluating direct financial impacts on stakeholders, particularly through the lens of stock market returns. Accordingly, the objective of this research is to analyze the impacts of the Mariana disaster on Vale’s stock returns using the synthetic control method. This methodology enables the construction of a transparent counterfactual scenario by combining weighted comparable units to estimate how Vale’s stock returns would have behaved in the absence of the disaster (Abadie & Gardeazabal, 2003 ; Abadie et al., 2010 ; Abadie et al., 2015 ). This design is especially suitable for settings with a single treated unit and multiple potential controls, because it formalizes the choice of the counterfactual as a data-driven weighted combination rather than relying on a single comparator (Maran, 2023 ). To strengthen identification and address potential spillovers, we analyze three donor pools: (i) a baseline pool including globally comparable firms indicated in Vale’s CVM Reference Form, the mandatory annual disclosure filing for Brazilian listed companies; (ii) a pool excluding Brazilian firms; and (iii) a stricter pool also excluding BHP Group Ltd. The analysis covers 2011 to 2017 and focuses on Mariana as the first major disaster directly associated with Vale within this window, reducing contamination from later shocks. This paper contributes in three ways. First, it adds evidence to disaster economics/finance by documenting how a major socio-environmental disaster is priced in equity markets (and for how long) using a flagship commodity producer as a revealing case. Second, it offers a structured synthetic-control design with alternative donor pools and robustness checks (including placebo-based inference), which is particularly useful in rare-event settings. Third, it derives implications for governance, operational risk management, and regulatory oversight in sectors where extreme events can have persistent valuation effects. Our findings indicate that Vale’s stock returns experienced an immediate decline following the disaster, although this initial drop was not as severe as projected by the counterfactual model. Nevertheless, from 2016.1 to 2017.2, Vale’s returns remained consistently below the expected values provided by the synthetic control, suggesting that the market initially underestimated the disaster’s longer-term impacts. Put differently, while the rupture was a discrete event, its consequences (environmental, financial, social, and legal) continued to shape valuation well beyond the initial shock. These results carry scientific and practical implications. Managers may infer the need for stronger investments in operational safety and sustainability, given the potential valuation penalties during and after extreme events. Investors should recognize that valuation approaches anchored in historical returns can understate “black swan” exposure, strengthening the case for more robust risk-management strategies. Regulators, in turn, may view the evidence as supporting enhanced transparency and preventive measures to strengthen market resilience in the face of rare, high-impact shocks like the Mariana disaster (Rhee & Wu, 2020 ). 2. Literature Review On November 5, 2015, the Fundão tailings dam (operated by Samarco, co-owned by Vale and BHP Billiton) failed in Mariana, Minas Gerais, abruptly releasing an estimated 55–62 million m³ of iron-ore tailings into the Doce River watershed. The slurry wave engulfed the district of Bento Rodrigues, displacing its residents and causing at least 19 fatalities, while contaminated sediments traveled approximately 663 km along the river system to the Atlantic coast. Beyond immediate destruction, the event posed persistent socioeconomic risks through impacts on water supply, fisheries, agriculture, and other ecosystem services in dozens of municipalities along the Doce River basin (Fernandes et al., 2016 ). From an economic standpoint, the disaster has also been treated as a long-lived regional shock rather than a short-lived disruption. Using a spatial panel difference-in-differences design, Batista and Firme ( 2025 ) find that direct economic losses were larger in municipalities closer to the Doce River and decreased with distance, with particularly severe effects in agriculture and industry. Their estimates suggest that direct losses accumulated substantially over the medium run, reaching approximately US $ 29.1 billion (after 3 years), US $ 57.1 billion (after 4 years), and US $ 95.5 billion (after 5 years) (in 2022 prices), while total net effects were smaller due to offsetting spillovers, approximately US $ 15.7 billion, US $ 28.0 billion, and US $ 49.1 billion over the same horizons (Batista & Firme, 2025 ). This broader context matters for a disaster-finance perspective: disasters may remain salient for years through litigation, remediation, regulation, and reputational pressure, channels that can reshape expectations about risk and future cash flows even after the “news shock” has passed. 2.1. Impacts of Disasters and Efficient Markets The mining sector, given its scale and its potential for severe environmental externalities, carries a non-trivial risk of operational disasters and, as a result, is frequently under public and regulatory scrutiny. As of 2023, Brazil had 927 registered dams, with 464 monitored under the National Dam Safety Policy (PNSB, Portuguese acronym). These structures are classified according to their Risk Category Index (CRI), underscoring the inherent operational risks associated with their management (ANM, 2023). Such risks became evident after the Fundão dam rupture in Mariana, widely recognized as one of the world’s most severe mining-related environmental disasters. Incidents of this magnitude expose vulnerabilities in governance, infrastructure integrity, and monitoring systems, reinforcing the need for stronger safety protocols and more transparent environmental reporting (Souza, 2019 ; Fabrício et al., 2021 ). They also highlight why environmental risks cannot be treated as “non-financial” in a narrow sense: disaster shocks may reshape expectations about cash flows, liabilities, and risk premia, with valuation consequences that extend beyond the immediate physical impact (Castro & Almeida, 2023 ). In other words, the disaster-economics literature stresses that disasters generate direct and indirect economic losses and can propagate through financial channels, affecting perceptions of risk and the pricing of affected entities (Maran, 2023 ). This discussion connects naturally to the Efficient Market Hypothesis (EMH). Fama ( 1970 ) argues that market prices incorporate available information, leaving little room for persistent abnormal returns. Under this view, new information should be rapidly absorbed by asset prices, making systematic outperformance unlikely. Yet rare and unpredictable events, often labeled “black swans”, challenge the EMH because they can be difficult to anticipate and may not be immediately priced, especially when uncertainty is high and information arrives in fragments (Taleb, 2007 ; Woods, 2006 ; Masys, 2012 ). Fama ( 1991 ) further notes that market expectations depend on a broader informational set, implying that assimilation may vary across contexts, particularly during unprecedented crises. Empirical evidence also suggests that corporate exposure and recovery are heterogeneous. Chang et al. ( 2022 ), for instance, show that disaster impacts depend on both internal factors (e.g., firm size, operational flexibility, reliance on local markets) and external conditions (e.g., regulation and government interventions). Insurance coverage and managerial capacity may shape how quickly firms stabilize and how investors revise expectations after shocks (Chang et al., 2022 ). From a public-finance angle, Chen ( 2020 ) highlights the government’s dual burden in crises, shrinking revenues alongside rising expenditures for response and restoration, an institutional setting that can amplify uncertainty around regulation, enforcement, and compensation. For economies with large commodity exporters, these mechanisms are particularly salient. Brazil has repeatedly faced environmental incidents that spilled over into corporate valuation debates, including Mariana (in 2015), Brumadinho (in 2019), and major oil spills involving Chevron (in 2011) and Petrobras (in 2019). The point is not merely that “prices fall” after bad news. Rather, disasters can trigger multi-year processes (remediation, litigation, regulatory revisions, and reputational reassessment) that affect perceived risk and expected cash flows. In that sense, the Mariana case is informative because it illustrates how disaster impacts can persist and diffuse through space and time at the regional level (Batista & Firme, 2025 ) and through ecological and socioeconomic channels that may alter livelihoods and local production conditions (Fernandes et al., 2016 ). In such settings, the central empirical challenge is not simply documenting price movements, but establishing a credible counterfactual: how would the firm have performed absent the disaster? This is precisely where synthetic control designs are useful. The method constructs a data-driven counterfactual as a weighted combination of comparable units, rather than relying on a single benchmark, and it is especially appropriate when there is one treated unit and many potential controls (Maran, 2023 ). Used carefully, it helps sharpen inference about whether observed post-event dynamics plausibly reflect disaster-related information and risk repricing. 2.2. Black Swans and Synthetic Control The Mariana disaster illustrates how large corporations can face abrupt market challenges when an extreme event unfolds and information diffuses rapidly. These shocks do not resemble typical macroeconomic fluctuations; they combine operational failure, environmental damage, and evolving legal and political responses. Taleb ( 2007 ) frames such episodes as “black swans”: rare, hard-to-predict events with outsized consequences across multiple domains. Long before that label became popular, economic and financial research emphasized that assets exposed to severe downside states may be discounted by risk-averse investors, while assets that remain resilient during crises become relatively more attractive (Roy, 1952 ; Menezes et al., 1980 ; Rietz, 1988 ; Rhee & Wu, 2020 ). Even though EMH and Black Swan Theory start from different premises (price efficiency under rational processing versus the dominance of the improbable), both speak to how economic agents respond to new information flows. The “black swan” metaphor is useful here mainly because it emphasizes model limits: rare events can invalidate comfortable inferences drawn from long periods of “normal times,” and crisis narratives may be constructed ex post in ways that appear coherent but do not necessarily reflect ex ante predictability (Taleb, 2007 ; Krupa & Jones, 2013 ). In practical terms, this tendency can lead managers and investors to underestimate tail risks, leaning too heavily on historical regularities that underweight low-probability, high-impact scenarios (Taleb, 2007 ; Krupa & Jones, 2013 ). In response to stakeholder pressure and growing awareness of environmental risk, firms in high-impact sectors have increasingly incorporated Corporate Social Responsibility (CSR) and broader sustainability frameworks. The shift toward a “triple bottom line” perspective, integrating social, environmental, and financial dimensions, also reflects an attempt to address investor and creditor concerns regarding long-run operational continuity and legitimacy (Martínez-Ferrero et al., 2015 ; Fabrício et al., 2021 ). Still, whether and how quickly markets incorporate these risks remains an empirical question, particularly when disasters trigger prolonged litigation, regulatory uncertainty, and reputational damage. A growing body of empirical work has examined the Mariana disaster and related events using diverse methods and outcomes. Simonato ( 2017 ) employs a computable general equilibrium approach and reports persistent negative regional impacts on production, employment, and trade. Castro and Almeida ( 2019 ) apply synthetic control methods to show significant losses in industrial and mining outputs in affected states after Mariana, highlighting the role of sectoral dependence. Carrillo et al. ( 2020 ), using a difference-in-differences design, identify adverse health effects among newborns exposed in utero, underscoring that disaster burdens may be unevenly distributed across social groups. Castro and Almeida ( 2023 ) extend the synthetic control framework to the Brumadinho disaster, documenting heterogeneous sectoral impacts and recovery patterns across events. More recently, Biazoli et al. ( 2024 ) combine clustering with synthetic control to build more robust counterfactuals, reinforcing the value of careful comparator construction in disaster settings. Relatedly, Kuruc ( 2022 ) illustrates synthetic control’s usefulness for evaluating crisis-related policy effectiveness, arguing that the approach can yield estimates that are less dependent on standard regression assumptions. Batista and Firme ( 2025 ) complement this evidence by explicitly modeling spatial spillovers and heterogeneous exposure across “neighbourhood levels”, documenting that losses were strongest closer to the Doce River and smaller farther away, while remaining economically meaningful in the medium run. Their results are helpful for a capital-markets lens because they underscore that Mariana was not a purely local and short-lived shock; rather, it generated a persistent and spatially structured economic footprint, a feature that can sustain investor attention and uncertainty over time (Batista & Firme, 2025 ). While much of this literature focuses on regional output, health, or sectoral activity, less attention has been given to isolating the disaster’s effect on the stock-market performance of a directly involved firm. This gap matters because market prices summarize expectations about cash flows, risk, and uncertainty, and they are a key channel through which disasters affect firms’ cost of capital and governance pressures. In this study, we therefore focus on Vale’s stock returns and use synthetic control methods to construct an explicit counterfactual path for returns in the absence of the Mariana shock. Given this framing, we articulate the following testable expectations: H1: Vale’s stock returns declined relative to the synthetic-control counterfactual immediately after the Mariana disaster. H2: Vale’s stock returns remained below the synthetic-control counterfactual in the post-disaster period, consistent with delayed market adjustment to the disaster’s longer-run implications. From these hypotheses, we develop an empirical strategy based on synthetic control methods, using comparable firms to build a robust counterfactual scenario. This design choice is aligned with recent contributions in the disaster-economics and disaster-finance literature that rely on transparent counterfactual construction when a single unit is exposed to a major shock (Maran, 2023 ). 2.3. Disaster-Economics and disaster-finance lens: channels to market pricing To connect the disaster-economics core of the journal to a capital-markets setting, it is useful to be explicit about the channels through which disasters can affect market pricing. First, disasters often generate large losses and sudden liquidity needs, while simultaneously increasing uncertainty about operational continuity and future liabilities. In settings where access to finance is limited or uneven, these frictions can shape recovery dynamics and the way losses translate into economic outcomes (Keerthiratne & Tol, 2017 ). This also raises an identification concern: a recurring issue in empirical disaster research is that common disaster measures can embed human and institutional components, complicating causal interpretation and strengthening the case for designs that emphasize credible counterfactuals and careful identification (Keerthiratne & Tol, 2017 ). Second, disaster-risk management is frequently framed as a component of resilience. Financial instruments (such as insurance, hedging strategies, and catastrophe-linked securities) are discussed in the literature as mechanisms that can transfer risk and smooth post-disaster fiscal or corporate stress (Halkos & Zisiadou, 2021 ). In this sense, market reactions to disasters can be interpreted as rapid updates in expectations about risk, financing conditions, and governance quality, not merely as responses to direct physical damage. Finally, these channels motivate why synthetic control is a useful design for disaster-finance questions. By producing an explicit counterfactual that does not depend on a single comparison unit, the method helps clarify whether observed valuation dynamics plausibly reflect disaster-related repricing rather than idiosyncratic contemporaneous shocks. Recent work has used synthetic control designs to evaluate disaster-related policy and financial mechanisms, precisely because the appeal lies in constructing a transparent counterfactual as a weighted combination of controls (Maran, 2023 ). This framing keeps the Vale case as an empirical trigger, while positioning the broader contribution as evidence on how environmental disasters may be (mis)priced (and for how long) in capital markets. 3. Method This section describes the empirical strategy used to evaluate how the Fundão tailings dam collapse (Mariana, Brazil; November 5, 2015) affected Vale S.A.’s stock returns, and whether the observed price dynamics are consistent with rapid versus delayed adjustment in the presence of a rare, high-impact shock. The treated unit is Vale, the outcome is the semiannual cumulative stock return, and the treatment occurs in 2015.2 (second semester of 2015). The analysis window spans 2011.2-2017.2, allowing us to compare Vale’s realized returns to a transparent counterfactual path built from comparable firms. Consistent with the hypotheses developed in the literature review, we interpret the estimated post-event gaps in two layers: (i) the immediate deviation in the disaster semester (H1), and (ii) the persistence of deviations in subsequent semesters (H2). 3.1. Construction of Synthetic Control for Vale To evaluate the impact of the Mariana disaster on the stock returns of Vale, it is essential to compare the company’s actual performance after the environmental incident with a hypothetical scenario in which the disaster did not occur. This hypothetical or counterfactual scenario cannot be directly observed. Thus, constructing a robust and reliable counterfactual becomes crucial for performing such comparisons. In this study, we employ the synthetic control method (SCM), originally proposed by Abadie and Gardeazabal ( 2003 ) and further developed by Abadie et al. ( 2010 , 2015 ). This technique is particularly suited for comparative case studies using aggregate data or situations involving few treated units, such as the specific case of Vale analyzed here. Unlike traditional methods like difference-in-differences, SCM emphasizes strong pre-treatment fit and an explicit counterfactual built from a weighted combination of controls, which reduces reliance on strong functional-form assumptions and helps clarify identification in settings where confounding and measurement challenges are common (Rosenbaum, 2007 ; Abadie et al., 2010 , 2015 ; Keerthiratne & Tol, 2017 ; Maran, 2023 ). A key step in this design is defining an appropriate donor pool and a predictor set so that the treated unit and the synthetic unit are as comparable as possible before the shock (Maran, 2023 ). The synthetic control method constructs a weighted average of selected control units (other companies with similar characteristics to Vale) to replicate the counterfactual trajectory of Vale’s stock returns. The weights attributed to these control units aim to minimize the Root Mean Squared Prediction Error (RMSPE) between the actual and synthetic trajectories before the disaster (Abadie et al., 2015 ). Consequently, the synthetic trajectory closely replicates the real trajectory observed before the event, enabling a meaningful analysis of the disaster’s impact. In operational terms, the estimated “effect” for each semester t is the gap between Vale’s observed return and the synthetic return: \(\:{\alpha\:}_{t}={SR}_{t}^{Vale}-{SR}_{t}^{Synth}\) ​ Where, \(\:{SR}_{t}^{Synth}\) is the weighted sum of donor-firm returns using the SCM weights. However, an essential assumption of the synthetic control method is the absence of spillover effects, where the event affects control units indirectly, potentially contaminating the counterfactual (Rosenbaum, 2007 ; Abadie et al., 2010 , 2015 ). In our analysis, spillover effects could manifest through impacts on iron ore prices, given Vale’s prominent global role. As Castro and Almeida ( 2023 ) highlight, disasters involving major commodity producers like Vale may significantly influence global commodity prices, thus affecting comparable companies and complicating the construction of the synthetic trajectory. In this setting, two macro-financial variables are particularly relevant: international iron ore prices and the U.S. dollar index, because they may affect mining firms’ returns. However, because these variables are common time shocks (i.e., they do not vary across firms in a given semester), they do not provide the cross-sectional variation required for the SCM optimization stage. For this reason, they were not included as predictors in the synthetic-control optimization, even though they remain economically meaningful drivers of returns in commodity-linked industries (Abadie et al., 2010 , 2015 ). In practice, the SCM design addresses such common shocks through pre-treatment matching and by comparing Vale to a counterfactual formed from firms exposed to the same global conditions. Vale operates globally, holding significant production, sales, and distribution activities across various countries. As illustrated in Fig. 1 , Vale is among the world’s leading iron ore producers, with a strong international market presence. In Brazil’s stock market, Vale is classified under Basic Materials in the “Metallic Minerals” segment. Alongside Vale, this segment includes only seven other companies: Aura Almas Mineração S.A., Aura Minerals Inc., Bradespar S.A., Companhia Brasileira de Alumínio, CSN Mineração S.A., Litel Participações S.A., and Litela Participações S.A. Given Vale’s significance in the global market and potential spillover risks, special attention was paid to selecting appropriate donor pools (control groups) to mitigate such risks. Following Castro and Almeida’s ( 2023 ) logic of mitigating contamination, while noting that their focus is regional economic outcomes and ours is firm-level stock returns, we adopted distinct donor pools to avoid contamination of the control units. The donor pool selection followed a two-step strategy. Initially, all competitor companies mentioned in Vale’s Reference Form were considered. Subsequently, due to data availability constraints, only companies present in the LSEG database were included, as detailed in Table 1 . Table 1 Quantity and selection criteria of Vale’s donor pools Description Quantity Rationale (brief) Companies cited in Vale’s Reference Form 27 Initial universe of peers disclosed by Vale (-) Companies not available in LSEG (11) Data availability constraint = Donor pool 1 (DP1) 16 Baseline eligible donor pool (-) Brazilian companies (3) Reduce potential domestic spillovers = Donor pool 2 (DP2) 13 Stricter donor pool (international only) (-) BHP Group Ltd (1) Avoid direct involvement with Samarco = Donor pool 3 (DP3) 12 Strictest donor pool Source: Research data (2026). The construction of three distinct donor pools was specifically designed to address potential spillover concerns systematically. DP1 comprises all identified companies initially available. DP2 excludes Brazilian firms (CSN, Gerdau, and Usiminas), as they might have experienced indirect spillover effects due to the disaster’s impact on the national market. DP3 further excludes BHP Group Ltd., a co-owner of Samarco (operator of the Fundão dam), to avoid potential direct contamination from spillover effects. The predictor variables utilized to construct the synthetic control were selected based on solid theoretical foundations and previous empirical studies, including Debt and Net Income (Sampaio, Azevedo & Azuaga, 2016 ). Additionally, to enhance the robustness of the synthetic control, firm size (market capitalization) and ESG controversies were included. The ESG Controversies Score is a scale of 0 to 100, measuring the company’s exposure to controversies/scandals and/or violations of ESG pillars captured by the media. A higher score indicates lower exposure to scandals. Therefore, this proxy provides an external perspective of the company (LSEG, 2025 ). The literature demonstrates a relationship between extreme environmental events and ESG disclosures and measures (Huang, Li, Lin, & McBrayer, 2022 ). The financial market itself ends up reacting negatively to companies that experience some type of scandal or suspicious activities, which impacts their market value (Aouadi & Marsat, 2018 ; Dalcero et al., 2024 ). Because SCM estimates can be sensitive to donor-pool composition and predictor choices, we report the three donor-pool constructions transparently and use them as structured sensitivity checks. This is consistent with methodological discussions emphasizing that robustness should be evaluated by examining whether estimated post-event gaps are driven by specific donor units or by the broader quality of the pre-treatment fit (Maran, 2023 ). Therefore, the meticulous construction of donor pools provides methodological robustness, enabling the analysis of Vale’s stock price responses to the Mariana disaster. This careful construction enhances the validity of the study, allowing us to evaluate the impact of rare and highly impactful incidents (Black Swan events) on corporate financial markets. 3.2. Data Collection and Analysis The observational units adopted in this study were Vale’s national and international competitor firms. We used stock returns as a proxy for companies’ economic and financial performance, allowing us to analyze how the financial market responded to the Mariana disaster. Given that the study included 16 companies from various countries: Australia (BHP Group Ltd, Fortescue Ltd, Rio Tinto Ltd), Brazil (Companhia Siderúrgica Nacional SA, Gerdau SA, Usinas Siderúrgicas de Minas Gerais SA), Canada (First Quantum Minerals Ltd), China (Jiangxi Copper Co Ltd), France (Eramet SA), Germany (Aurubis AG), South Africa (Kumba Iron Ore Ltd), Switzerland (Ferrexpo PLC, Glencore PLC), United Kingdom (Anglo American PLC, Antofagasta PLC), and the United States (Freeport-McMoRan Inc). We needed to adjust accounting data to accommodate the shortest available reporting interval in the database. Hence, stock returns were adjusted to a semiannual basis, considering that international companies in our sample report their financial statements semiannually in the LSEG database. Throughout the paper, “.1” denotes the first semester (January-June) and “.2” denotes the second semester (July-December). Previous studies that examined the environmental disaster’s impact on Vale’s returns primarily adopted the event study methodology, focusing on cumulative returns around the event date (Fogaça, Raeder & Marques, 2023 ; Rocha & Vasconcelos, 2023 ). Although our research employs a different approach, it is important to acknowledge that the broader temporal aggregation we use can limit precision in capturing very short-lived immediate effects. However, this broader scope is well aligned with our objective of investigating whether disaster-related repricing persists beyond the initial shock, which is central to testing delayed adjustment in the semesters following the event. Monthly returns for the analysis period were directly extracted from the LSEG Eikon Datastream. Once obtained, we calculated cumulative returns for each semester. The use of stock returns as the main variable is grounded in the financial literature based on the Efficient Market Hypothesis (EMH), which posits that financial asset prices reflect all available information, both public and private (Fama, 1970 , 1991 ). Additionally, this variable is widely employed in research examining the financial market impacts of adverse events (Fogaça, Raeder & Marques, 2023 ; Rocha & Vasconcelos, 2023 ). Additional information about the variables used can be found in Table 2 . Table 2 Description of variables used Variable Abbreviation Type Description Source Stock Return SR Dependent Semiannual cumulative return = \(\:{\prod\:}_{m=1}^{6}\left(1+{RM}_{m}\right)\) -1 LSEG Eikon Firm Size SIZE Predictor Proxy used was Company Market Capitalization LSEG Eikon Debt DEB Predictor (Short-term and long-term interest-bearing liabilities) / Total Assets LSEG Eikon ESG Controversies ESG Predictor Methodology of the database itself LSEG Eikon Net Income NI Predictor Proxy used was Net Income LSEG Eikon Source: Research data (2026). The predictor variables derived from company financial statements were aggregated semiannually, which corresponds to the smallest available time frame for the entire sample. The analysis covers the period from the second half of 2011 to the second half of 2017, comprising 13 semesters. Given that the Fundão dam collapse in Mariana occurred on November 5, 2015, the semesters studied were divided as follows: 8 semesters pre-disaster (2011.2 to 2015.1), 1 semester during the disaster (2015.2), and 4 semesters post-disaster (2016.1 to 2017.2). The starting period was chosen due to data availability for all analyzed companies (e.g., Glencore PLC started trading in the first semester of 2011, making its data available from 2011.2 onwards). The endpoint of our analysis was a discretionary choice that does not compromise our findings. According to Abadie et al. ( 2010 , 2015 ), the synthetic control technique yields optimal results when the number of observations before the event is larger than the number of observations after it. Finally, to connect the empirical design to the hypotheses, we interpret the post-event gaps \(\:{\alpha\:}_{t}\) as follows. H1 is assessed by examining the sign and magnitude of the gap in 2015.2 (the disaster semester). H2 is assessed by examining whether the gaps remain negative in the post-disaster semesters (2016.1-2017.2), consistent with persistent underperformance relative to the counterfactual. In line with standard SCM practice, inference is supported by transparency in pre-treatment fit (RMSPE) and by placebo-style comparisons that evaluate whether the estimated post-event path for Vale is unusual relative to gaps obtained when the “treatment” is reassigned to donor units (Abadie et al., 2010 , 2015 ). These procedures complement the structured sensitivity analysis provided by DP1-DP3 and help assess whether the estimated disaster-related deviation is plausibly distinct from the distribution of placebo gaps. 4. Results Analysis This section reports and interprets the synthetic-control estimates for Vale using three donor pools: DP1 (all firms in the final sample), DP2 (excluding Brazilian firms to mitigate potential domestic spillovers), and DP3 (excluding both Brazilian firms and BHP Group Ltd., given its direct link to Samarco at the time of the disaster). The goal is to test our two expectations from the literature: (H1) an immediate negative gap around the disaster semester (2015.2), and (H2) a persistent post-disaster gap consistent with delayed market adjustment. Table 3 summarizes the donor-pool composition and the resulting pre-disaster fit (RMSPE) for each specification. Table 3 Companies’ contribution to weight formation in DP1, DP2, and DP3 Companies Country Ticker DP1 Weight DP2 Weight DP3 Weight Anglo American PLC United Kingdom AAL.L 0.049 0.064 - Antofagasta PLC United Kingdom ANTO.L 0.028 0.039 - Aurubis AG Germany NAFG.DE 0.033 0.044 - BHP Group Ltd Australia BHP.AX 0.315 0.315 X Companhia Siderúrgica Nacional SA Brazil CSNA3.SA 0.009 X X Eramet SA France ERMT.PA 0.036 0.047 - Ferrexpo PLC Switzerland FXPO.L 0.012 0.020 - First Quantum Minerals Ltd Canada FM.TO 0.040 0.054 - Fortescue Ltd Australia FMG.AX 0.012 0.021 0.064 Freeport-McMoRan Inc United States FCX.N 0.031 0.046 - Gerdau SA Brazil GGBR4.SA 0.025 X X Glencore PLC Switzerland GLEN.L 0.121 0.109 - Jiangxi Copper Co Ltd China 600362.SS 0.076 0.038 - Kumba Iron Ore Ltd South Africa KIOJ.J 0.035 0.045 0.147 Rio Tinto Ltd Australia RIO.L 0.159 0.158 0.789 Usinas Siderurgicas de Minas Gerais SA Brazil USIM5.SA 0.019 X X RMSPE 0.00916 0.00890 0.00561 Note: RMSPE is the Root Mean Squared Prediction Error; “X” means the company was excluded from the donor pool; “-” indicates the company was available but not selected for synthetic control. The sum of the weights in each column equals 1 (100%). Source: Research data (2026). Table 3 indicates that DP1 provides the best pre-disaster fit, with the lowest RMSPE (0.00916). DP1 also yields a broadly distributed synthetic composition, with meaningful contributions from several global peers. In DP1, the largest weights are assigned to BHP Group (0.315), Rio Tinto (0.159), and Glencore (0.121), followed by Jiangxi Copper (0.076). This pattern is consistent with the logic of SCM: the synthetic unit is constructed as a weighted combination of controls that best matches the treated unit in the pre-event period (Abadie et al., 2010 , 2015 ). Importantly, this is also the reason donor-pool definition and pre-treatment comparability are central design choices in SCM applications (Maran, 2023 ). DP2, which excludes Brazilian firms to reduce potential national-market spillovers, delivers a weaker (though still acceptable) pre-disaster fit (RMSPE 0.00890). The weighting structure remains diversified, with sizable contributions from BHP (0.315), Rio Tinto (0.158), and Glencore (0.109), plus moderate weights spread across other international peers. In substantive terms, DP2 functions as a spillover-robustness check: it reduces the likelihood that domestic firms’ co-movement with Vale (driven by shared Brazilian macro-financial conditions) mechanically improves fit. DP3 is the most restrictive specification, excluding Brazilian firms and BHP Group Ltd. Given BHP’s direct association with Samarco, this exclusion is conceptually appealing from a contamination standpoint, but it comes with a clear trade-off: DP3 delivers the poorest pre-disaster fit (RMSPE 0.00561) and produces a highly concentrated synthetic control. Only three firms receive positive weights, dominated by Rio Tinto (0.789), followed by Kumba Iron Ore (0.147) and Fortescue (0.064). This concentration heightens sensitivity to idiosyncratic features of a single comparator, an issue often discussed as a practical limitation of restrictive donor pools in SCM designs (Abadie et al., 2010 ; Maran, 2023 ). For that reason, we treat DP3 primarily as a robustness bound, rather than as the preferred baseline. To complement Table 3 , Table 4 summarizes (i) predictor balance between Vale and its synthetic counterpart (illustrative for DP1), and (ii) the relative importance (“predictor weights”) assigned during optimization. For readability and replication, monetary values are reported in US $ billions (rounded), while ratios and scores retain their natural scales. Table 4 Contribution of variables to weight formation in DP1, DP2, and DP3 Predictor Treated (Vale) Synthetic (DP1) Sample Mean DP1 Weight DP2 Weight DP3 Weight Firm Size (Market Cap, US $ bn) 76.33 76.33 28.55 0.130 - 0.540 Debt (Debt/Total Assets) 0.229 0.229 0.261 0.405 0.479 0.458 Net Income (US $ bn) 1.900 1.900 0.520 0.093 0.027 0.002 ESG Controversies Score 50.04 50.05 77.66 0.372 0.493 - RMSPE 0.00916 0.00890 0.00561 Note: “-” indicates the predictor was available but not selected (i.e., received zero weight) in that donor-pool optimization. The sum of the weights in each column equals 1 (100%). Source: Research data (2026). Across specifications, Debt receives the largest (or near-largest) predictor weight (DP1 = 0.405; DP2 = 0.479; DP3 = 0.458), suggesting that capital-structure similarity is pivotal for reproducing Vale’s pre-disaster return trajectory. This is economically coherent in a disaster-finance setting: leverage is a salient risk indicator when shocks raise uncertainty about liabilities, cash flows, and financing conditions. In parallel, the ESG Controversies Score receives substantial weight in DP1 (0.372) and DP2 (0.493), consistent with the view that scandals/controversies can be material for valuation and can interact with how markets process disaster-related news (Huang et al., 2022 ; Aouadi & Marsat, 2018 ; Dalcero et al., 2024 ). The prominence of these predictors supports a disaster-finance interpretation in which the post-event gap may reflect repricing of risk, financing conditions, and perceived governance quality after the shock (Maran, 2023 ; Halkos & Zisiadou, 2021 ). Having established donor-pool fit and predictor structure, we now turn to the estimated effects in the event semester and post-event period. Figures 2 – 4 plot Vale’s observed returns against the synthetic counterfactual for each donor pool. For consistency with the two hypotheses, we discuss results in two steps: (i) the disaster semester (2015.2), aligned with H1, and (ii) the post-disaster semesters (2016.1-2017.2), aligned with H2. Source: Prepared by the authors (2026). In DP1, the pre-disaster period (2011.2-2015.1) shows a close alignment between Vale and its synthetic counterpart, consistent with the extremely low RMSPE reported in Table 3 . Minor deviations are expected even under strong fit and do not undermine identification, as SCM inference hinges on the overall quality of the pre-event match (Abadie et al., 2010 , 2015 ). Testing H1 (immediate effect in 2015.2) : in the disaster semester (2015.2), Vale’s observed returns decline, indicating an immediate negative response aligned with the idea that severe shocks can be capitalized into prices. Interestingly, the synthetic series shows an even sharper decline in the same semester. Rather than treating this as a contradiction, it is more informative to interpret it as evidence that immediate pricing may be shaped by offsetting forces: while disaster news increases uncertainty and perceived tail risk, short-run price dynamics may also embed expectations about remediation, commodity-market conditions, firm prominence, and investors’ prior beliefs. From a disaster-finance lens, what matters is that disasters can trigger abrupt updates in perceived downside risk and liquidity/financing constraints, channels that can appear with lags or be partially masked in the immediate window (Maran, 2023 ). Thus, H1 is supported in terms of sign (returns fall), but the DP1 comparison suggests that the immediate “gap” is not the main margin of adjustment in this case. Testing H2 (persistence after 2015.2) : the more pronounced and economically meaningful result emerges in the post-disaster period. From 2016.1 onward, Vale’s returns remain systematically below the synthetic counterfactual for an extended sequence of semesters. This pattern is consistent with delayed adjustment: markets may underreact initially when the long-run scope of legal, regulatory, and reputational consequences is difficult to quantify, and then gradually incorporate disaster-related information as uncertainties resolve. This is precisely the type of mechanism emphasized in disaster-economics/finance research: beyond immediate physical losses, disasters can propagate through financial channels and reshape risk premia and financing conditions over time (Maran, 2023 ). In the Mariana context, prolonged litigation and compensation uncertainty are plausible channels sustaining risk salience, which can translate into persistent valuation discounts. Although DP1 is the baseline due to superior pre-event fit, DP2 and DP3 provide useful robustness checks. In DP2, the pre-event tracking remains visually close (though weaker than DP1 in RMSPE terms), and the post-event pattern continues to suggest that Vale underperforms its counterfactual path over multiple semesters. In DP3, the comparison should be interpreted with greater caution, given its high RMSPE and heavy reliance on a single firm (Rio Tinto). Even so, DP3 is directionally consistent with the broader narrative: the post-event period exhibits a sustained divergence that is difficult to reconcile with instantaneous and complete adjustment. Taken together, these comparisons strengthen the inference that the main empirical phenomenon is persistence, not merely an immediate dip. Figures 3 and 4 present robustness analyses with DP2 and DP3. Source: Prepared by the authors (2026). Source: Prepared by the authors (2026). To assess whether the estimated post-event divergence could be driven by idiosyncratic noise rather than the disaster shock, we performed a placebo test in which each donor-pool firm (in DP1) is iteratively treated as if it experienced the Mariana disaster. This produces a distribution of placebo gaps that serves as an informal benchmark for the “unusualness” of Vale’s estimated effect (Abadie et al., 2010 , 2015 ). Figure 5 summarizes these placebo trajectories. Source: Prepared by the authors (2026). Before 2015.2, placebo trajectories tend to fluctuate around zero with limited dispersion, which is expected if pre-event fit is broadly adequate across units. After 2015.2, dispersion widens, reflecting heterogeneity in firms’ return dynamics and exposure to global commodity-cycle movements and firm-specific shocks. The key interpretive step is not simply that “dispersion exists,” but whether Vale’s post-event gap is large relative to the placebo distribution, especially when benchmarked against pre-event fit. In SCM practice, this logic is often strengthened by comparing post-to-pre RMSPE ratios (or related normalized metrics), which helps distinguish large effects from artifacts of poor fit (Abadie et al., 2010 , 2015 ). In our setting (where DP1 fit is extremely strong), the sustained negative divergence for Vale in the post-disaster period is consistent with a meaningful disaster-related repricing rather than random fluctuation. Two implications follow from the pattern of results. First, the results suggest that the Mariana disaster was not fully priced as a one-off “news shock”. Instead, pricing appears to incorporate the disaster’s consequences more gradually, consistent with the idea that disasters can create persistent financial frictions, through uncertainty about liabilities, governance and compliance expectations, and the perceived reliability of risk controls (Maran, 2023 ). Second, the evidence speaks directly to governance and resilience: if markets penalize firms for prolonged periods after major socio-environmental disasters, then investments in disaster-risk management, preparedness, and transparency can be interpreted not only as ethical or regulatory priorities but also as strategies to reduce valuation losses under tail risks (Halkos & Zisiadou, 2021 ). In capital markets, where prices aggregate beliefs about risk and future cash flows, these mechanisms can translate into sustained discounts, particularly in high-impact sectors where operational failures are tightly linked to regulatory scrutiny and stakeholder pressure. Overall, the results align with our framing in the Introduction and Literature Review: H1 is supported in the sense that returns decline in the disaster semester, but the central evidence is the persistent post-event gap consistent with H2, suggesting delayed market adjustment following a rare, high-impact socio-environmental shock. 5. Conclusion The evidence from this study points to a clear takeaway: in a major socio-environmental disaster, the most consequential valuation effects may arise not only from the initial shock but from the persistence of uncertainty and risk repricing over subsequent periods. For Mariana, the results indicate a short-run drop in Vale’s returns in the disaster semester (2015.2), followed by a more informative pattern: a sustained underperformance relative to a credible counterfactual in 2016.1-2017.2. This dynamic is difficult to reconcile with strict interpretations of instantaneous, full incorporation of disaster-related information, and instead aligns with delayed price adjustment under a rare, high-impact event. The immediate semester of the rupture provides an important nuance. Vale’s return decline in 2015.2 is consistent with an adverse market reaction (H1), yet the drop is not as severe as the synthetic benchmark would suggest. Rather than treating this as a paradox, a more plausible reading is that early pricing combined multiple, partially offsetting signals: the shock itself, expectations about compensation and remediation, and uncertainty about the timing and scale of liabilities. In disaster contexts, information rarely arrives as a single “clean” announcement (legal developments, enforcement, operational constraints, and reputational damage often unfold in fragments). Under such conditions, initial returns can reflect a provisional narrative that is later revised. The post-disaster trajectory is therefore the core result. The persistent gap between Vale and the synthetic counterfactual (H2) suggests that the market initially underweighted longer-run consequences and gradually incorporated them as uncertainty evolved. From a disaster-finance lens, this pattern is economically meaningful because it is consistent with repricing of perceived downside risk, liquidity pressure, and financing conditions, channels through which disasters can affect valuation beyond direct physical losses (Maran, 2023 ). In practical terms, the market appears to have treated Mariana not merely as a one-off accident but as the beginning of a multi-period process with governance, legal, and operational implications. This paper contributes to the literature in three ways, and each contribution is anchored in what the results can (and cannot) support. First, it offers firm-level evidence on how an extreme socio-environmental disaster is (mis)priced over time in equity markets. Much of the existing Mariana-related evidence emphasizes regional economic activity, sectoral outcomes, or health impacts; our focus complements those perspectives by isolating a stakeholder-relevant capital-markets outcome and tracing its persistence. Second, the paper provides a structured synthetic-control design that makes donor-pool sensitivity visible rather than implicit. By contrasting DP1, DP2, and DP3, the study shows how spillover concerns and comparability trade-offs can change pre-fit and inference, which is precisely the type of transparency expected in applied disaster-economics work. Third, the placebo-style exercise supports the interpretation that the post-event divergence is not merely noise, reinforcing the credibility of the counterfactual-based narrative when pre-event fit is strong. The implications follow directly from the persistence result. For investors, the evidence cautions against reading disaster risk primarily through short event windows or “first-day” market reactions. When disasters trigger prolonged litigation, regulatory adjustments, and reputational reassessment, early price movements may understate the longer-run valuation penalty. This matters for portfolio construction, for scenario analysis, and for any valuation approach that extrapolates from historical returns while underweighting tail risk. For managers, especially in environmentally sensitive sectors, our findings underscore that disaster-risk management is not a peripheral compliance item. It can be interpreted as a value-preserving component of resilience that shapes how shocks translate into perceived risk and expected cash flows. Are operational safety systems, disclosure practices, and governance arrangements designed to remain credible under stress, or only under normal times? The broader disaster-finance literature highlights that preparedness and financial mechanisms (including insurance and related instruments) may influence how shocks propagate through corporate financial stress and market expectations, reinforcing the relevance of integrating disaster-risk thinking into governance and financial planning (Halkos & Zisiadou, 2021 ). For regulators and policymakers, the results support a concrete rationale for stronger oversight and more credible disclosure regimes around operational and socio-environmental risks. If markets may underreact initially and adjust gradually as uncertainty resolves, then improving transparency and enforcement can reduce informational frictions that prolong valuation discounts and weaken trust. In tailings dam contexts, this points to the importance of monitoring systems, clear liability frameworks, and disclosure standards that allow stakeholders to assess both near-term exposure and longer-run contingent risks. This study has limitations that also define a feasible agenda for future research. The sample window is designed to preserve a strong pre-treatment fit and minimize contamination from later shocks, but it may not capture effects that extend further into the long run. In addition, the predictor set is constrained by database coverage and may not fully reflect all dimensions relevant for return dynamics (e.g., richer governance proxies or alternative risk measures). Future work can extend the horizon, explore complementary outcomes (volatility, downside risk, cost-of-capital proxies), and compare Mariana with other disasters, to test whether delayed adjustment is a recurring feature of socio-environmental shocks or varies with institutional response, media salience, and litigation dynamics. In sum, the main message is specific: the Mariana disaster is associated with an immediate negative return shock, but the more informative evidence lies in the persistent post-disaster gap relative to a credible counterfactual, consistent with gradual repricing of risk and uncertainty rather than instantaneous incorporation. This is the “so what” for disaster finance: governance quality, transparency, and risk management can plausibly influence not only the size of the initial valuation impact, but also how long the market continues to discount the firm after an extreme event. Declarations Competing interests The authors declare that they have no competing interests. Funding The authors received no specific funding for this work. Author Contribution Matheus Soares Mendes, João Antonio Da Costa Neto, Francisco De Assis Miranda Da Silva, and Orleans Silva Martins jointly conceived and designed the research project. All authors contributed equally to the development of the research question, theoretical framing, and formulation of hypotheses. They collaboratively conducted the literature review, designed the empirical strategy, and defined the methodological procedures. Data collection, organization, and analysis were performed in a coordinated manner, with all authors actively participating in the interpretation of results and discussion of implications. The manuscript was written collectively, with iterative rounds of drafting and revision involving all authors. All authors critically reviewed the final version, approved it for submission, and agree to be accountable for all aspects of the work. Data Availability The data that support the findings of this study are available from third-party providers (e.g., LSEG) under license restrictions. Replication code and derived data used in the analysis can be made available by the corresponding author upon reasonable request. References Abadie A, Diamond A, Hainmueller J (2010) Synthetic control methods for comparative case studies: estimating the effect of California’s tobacco control program. J Am Stat Assoc 105(490):493–505. https://doi.org/10.1198/jasa.2009.ap08746 Abadie A, Diamond A, Hainmueller J (2015) Comparative politics and the synthetic control method. Am J Polit Sci 59(2):495–510. https://doi.org/10.1111/ajps.12116 Abadie A, Gardeazabal J (2003) The economic costs of conflict: a case study of the Basque Country. Am Econ Rev 93(1):113–132. https://doi.org/10.1257/000282803321455188 Relatório anual de segurança de barragens de mineração 2023. ANM, ANM - Agência Nacional de Mineração, Brasília (2023) Available at: https://www.gov.br/anm/pt-br/assuntos/barragens/relatorios-anuais-de-seguranca-da-barragens-de-mineracao-2/relatorio-anual-2023-final-1.pdf Aouadi A, Marsat S (2018) Do ESG controversies matter for firm value? Evidence from international data. J Bus Ethics 151:1027–1047. https://doi.org/10.1007/s10551-016-3213-8 Batista ICAA, Firme VAC (2025) Economic impacts of Brazil’s Mariana dam disaster on different levels of neighbourhood. Spat Economic Anal 20(4):728–750. https://doi.org/10.1080/17421772.2025.2502355 Biazoli L, de Ávila ES, de Oliveira IRC (2024) Combining cluster analysis with synthetic control for evaluating economic impacts of the dam breach in Mariana, Brazil. https://doi.org/10.1007/s00181-024-02627-7 . Empirical Economics Carrillo B, da Mata D, Emanuel L, Lopes D, Sampaio B (2020) Avoidable environmental disasters and infant health: evidence from a mining dam collapse in Brazil. Health Econ 29(12):1–9. https://doi.org/10.1002/hec.4151 Castro LSD, Almeida ES (2019) Desastres e desempenho econômico: avaliação do impacto do rompimento da Barragem de Mariana. Geosul 34(70):406–429. https://doi.org/10.5007/2177-5230.2019v34n70p406 Castro LSD, Almeida E (2023) Avaliação do desastre de Brumadinho no desempenho econômico de Minas Gerais. Nova Economia 33(2):421–447. https://doi.org/10.1590/0103-6351/7658 Chang SE, Brown C, Handmer J, Helgeson J, Kajitani Y, Keating A, Roa-Henriquez A (2022) Business recovery from disasters: lessons from natural hazards and the COVID-19 pandemic. Int J Disaster Risk Reduct 80:103191. https://doi.org/10.1016/j.ijdrr.2022.103191 Chen G (2020) Assessing the financial impact of natural disasters on local governments. Public Budg Finance 40(1):22–44. https://doi.org/10.1111/pbaf.12245 Dalcero K, Rover S, Ferreira DDM (2024) Um mar de lama: há impacto no valor de mercado da Vale S/A pós desastres? Revista de Administração. Contabilidade e Economia da Fundace 15(2). https://doi.org/10.13059/racef.v15i2.1181 Duarte AP (2008) Classificação das barragens de contenção de rejeitos da mineração e de resíduos industriais no estado de Minas Gerais em relação ao potencial de risco. MSc dissertation, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil Fabrício SA, Ferreira DDM, Borba JA (2021) A panorama of Mariana and Brumadinho disasters: what do we know so far? REAd. Revista Eletrônica de Administração 27:128–152. https://doi.org/10.1590/1413-2311.310.102806 Fama EF (1970) Efficient capital markets: a review of theory and empirical work. J Finance 25(2):383–417 Fama EF (1991) Efficient capital markets: II. J Finance 46(5):1575–1617 Fernandes GW, Goulart FF, Ranieri BD et al (2016) Deep into the mud: ecological and socio-economic impacts of the dam breach in Mariana, Brazil. Natureza & Conservação 14(2):35–45. https://doi.org/10.1016/j.ncon.2016.10.003 Fogaça PACS, Raeder FT, Marques JAVC (2023) Análise dos impactos dos acidentes ambientais de Mariana e Brumadinho nas ações da Mineradora Vale. REUNIR: Revista de Administração, Contabilidade e Sustentabilidade 13(2):1–18. https://doi.org/10.18696/reunir.v13i2.1426 Halkos G, Zisiadou A (2021) Can we hedge an investment against a potential unexpected environmental disaster? Econ Disasters Clim Change 5:355–365. https://doi.org/10.1007/s41885-021-00085-4 Huang Q, Li Y, Lin M, McBrayer GA (2022) Natural disasters, risk salience, and corporate ESG disclosure. J Corp Finance 72:102152. https://doi.org/10.1016/j.jcorpfin.2021.102152 IBRAM - Instituto Brasileiro de Mineração (2024) Contribuição da indústria da mineração para o G20: construindo um mundo justo e um planeta sustentável. IBRAM, Brasília. Available at: https://ibram.org.br/wp-content/uploads/2024/11/IBRAM_Contribuicao-da-Industria-da-Industria-da-Mineracao-para-o-G20-Brasil-2024-2-1.pdf Keerthiratne S, Tol RSJ (2017) Impact of natural disasters on financial development. Econ Disasters Clim Change 1:33–54. https://doi.org/10.1007/s41885-017-0002-5 Krupa J, Jones C (2013) Black swan theory: applications to energy market histories and technologies. Energy Strategy Reviews 1(4):286–290. https://doi.org/10.1016/j.esr.2013.02.004 Kuruc K (2022) Are IMF rescue packages effective? A synthetic control analysis of macroeconomic crises. J Monet Econ 127:38–53. https://doi.org/10.1016/j.jmoneco.2022.02.002 LSEG (2025) LSEG ESG Scores: methodology. Available at: https://www.refinitiv.com/content/dam/marketing/en_us/documents/methodology/esg-scores-methodology.pdf (accessed 12 Dec 2025) Maran R (2023) Do sovereign catastrophe bonds improve fiscal resilience? An application of synthetic control method to Mexico (preprint). https://doi.org/10.21203/rs.3.rs-2965725/v1 . Research Square Mansur R (2025) Mariana, 10 anos: ninguém foi condenado por tragédia que matou 19 pessoas, destruiu comunidades e contaminou Rio Doce. G1 (3 Nov 2025). Available at: https://g1.globo.com/mg/minas-gerais/noticia/2025/11/03/mariana-10-anos-ninguem-foi-condenado-por-tragedia-que-matou-19-pessoas-destruiu-comunidades-e-contaminou-rio-doce.ghtml Martínez-Ferrero J, García-Sánchez IM, Cuadrado-Ballesteros B (2015) Effect of financial reporting quality on sustainability information disclosure. Corp Soc Responsib Environ Manag 22(1):45–64. https://doi.org/10.1002/csr.1330 Masys AJ (2012) Black swans to grey swans: revealing the uncertainty. Disaster Prev Manage 21(3):320–335. https://doi.org/10.1108/09653561211234507 Menezes C, Geiss C, Tressler J (1980) Increasing downside risk. Am Econ Rev 70:921–932 Rhee SG, Wu FH (2020) Conditional extreme risk, black swan hedging, and asset prices. J Empir Finance 58:412–435. https://doi.org/10.1016/j.jempfin.2020.07.002 Rietz T (1988) The equity risk premium: a solution. J Monet Econ 22(1):117–131 Rocha CH, Vasconcelos JG (2023) Barragens de Mariana e Brumadinho e preço da ação da Vale: um estudo de evento. Revista Estudos e Pesquisas em Administração 7(1). https://doi.org/10.30781/repad.v7i1.14785 Rosenbaum PR (2007) Interference between units in randomized experiments. J Am Stat Assoc 102(477):191–200. https://doi.org/10.1198/016214506000001112 Roy AD (1952) Safety first and the holding of assets. Econometrica 20:431–449 Sampaio B, Azevedo R, Azuaga FL (2016) Impacto de mudança regulatória sobre os preços dos ativos das empresas do setor elétrico brasileiro. Anais do 44º Encontro Nacional de Economia, Foz do Iguaçu, PR, Brazil Simonato TC (2017) Projeção dos impactos econômicos regionais do desastre de Mariana-MG. MSc dissertation, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil Souza PA (2019) Open data could have helped us learn from another mining dam disaster. Sci Data 6:54 Taleb NN (2007) The Black Swan: the impact of the highly improbable. Penguin Books, London WMTF - World Mine Tailings Failures (2019) World mine tailings failures (database). Available at: https://worldminetailingsfailures.org (accessed 1 Nov 2024) Woods DD (2006) Essential characteristics of resilience. In: Hollnagel E, Woods DD, Leveson N (eds) Resilience engineering: concepts and precepts. Ashgate, Aldershot, pp 21–34 Footnotes Approximately US $ 27.5 billion, converted at the year-end 2024 PTAX rate of R $ 6.1923 per US $ 1. Original values (R $ of 2022): direct losses R $ 151.9 bn, R $ 297.9 bn, R $ 498.2 bn; total net effects R $ 81.7 bn, R $ 146.3 bn, R $ 256.0 bn (Batista & Firme, 2025 ) Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 23 Apr, 2026 Reviews received at journal 21 Apr, 2026 Reviews received at journal 06 Apr, 2026 Reviewers agreed at journal 17 Mar, 2026 Reviewers agreed at journal 03 Mar, 2026 Reviewers invited by journal 03 Mar, 2026 Editor assigned by journal 23 Feb, 2026 Submission checks completed at journal 19 Feb, 2026 First submitted to journal 16 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8894226","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":601081879,"identity":"5949b4e3-eecf-4285-a48f-46f169f233b8","order_by":0,"name":"Matheus Soares Mendes","email":"","orcid":"","institution":"Federal University of Paraíba","correspondingAuthor":false,"prefix":"","firstName":"Matheus","middleName":"Soares","lastName":"Mendes","suffix":""},{"id":601081880,"identity":"d9bb56cd-e61c-4818-af4f-16aad5e2a6b8","order_by":1,"name":"João Antônio da Costa Neto","email":"","orcid":"","institution":"Federal University of Paraíba","correspondingAuthor":false,"prefix":"","firstName":"João","middleName":"Antônio da Costa","lastName":"Neto","suffix":""},{"id":601081881,"identity":"50dd8512-1f98-4e7e-aee7-a3fc0aa895d8","order_by":2,"name":"Francisco de Assis Miranda da Silva","email":"","orcid":"","institution":"Federal University of Paraíba","correspondingAuthor":false,"prefix":"","firstName":"Francisco","middleName":"de Assis Miranda da","lastName":"Silva","suffix":""},{"id":601081882,"identity":"266d6135-f431-4275-ac31-70ea6e56be2c","order_by":3,"name":"Orleans Silva Martins","email":"data:image/png;base64,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","orcid":"","institution":"Federal University of Paraíba","correspondingAuthor":true,"prefix":"","firstName":"Orleans","middleName":"Silva","lastName":"Martins","suffix":""}],"badges":[],"createdAt":"2026-02-16 15:08:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8894226/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8894226/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104055178,"identity":"a1b047af-963e-4b53-8ba3-c5ceef7b87c1","added_by":"auto","created_at":"2026-03-06 08:30:24","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":124379,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVale’s global operations and markets\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSource: Vale Reference Form (2023).\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8894226/v1/ee29a0a3d514356ff7025349.jpg"},{"id":104055179,"identity":"a97fa190-6156-4e4d-85c7-cebab3dccea9","added_by":"auto","created_at":"2026-03-06 08:30:24","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":96418,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImpact of the Mariana disaster on Vale’s stock returns (DP1)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSource: Prepared by the authors (2026).\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8894226/v1/c2458a2c94f92bd498bc7464.jpg"},{"id":104055177,"identity":"b9dc43f4-d654-4a1a-83b6-cf03e05fe93c","added_by":"auto","created_at":"2026-03-06 08:30:24","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":91718,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImpact of the Mariana disaster on Vale’s stock returns (DP2)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSource: Prepared by the authors (2026).\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8894226/v1/881afd9f9c661ededa3a20c3.jpg"},{"id":104055175,"identity":"a99f15c6-99bd-4650-9708-0602f5574ec5","added_by":"auto","created_at":"2026-03-06 08:30:24","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":94076,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImpact of the Mariana disaster on Vale’s stock returns (DP3)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSource: Prepared by the authors (2026).\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8894226/v1/979911e8aac758f8decdecf9.jpg"},{"id":104779400,"identity":"87be91cc-2bf6-4d4a-ad4c-117264449f37","added_by":"auto","created_at":"2026-03-17 07:39:49","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":124243,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePlacebo test\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSource: Prepared by the authors (2026).\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8894226/v1/0e3fbcc609f57e61d6e385ab.jpg"},{"id":104787283,"identity":"45ff6117-2b12-49cb-8a57-075339995271","added_by":"auto","created_at":"2026-03-17 08:20:22","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1569212,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8894226/v1/302044b6-803d-433d-afb0-9329da389cca.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"How Do Equity Markets Price Disaster Risk? Evidence from Vale after the Mariana Disaster","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eA decade after the collapse of the Fund\u0026atilde;o dam, the Mariana disaster remains a critical milestone in Brazil\u0026rsquo;s mining history, given the persistent legal and social challenges. Although a landmark R\u003cspan\u003e$\u003c/span\u003e 170\u0026nbsp;billion (\u0026asymp;\u0026thinsp;US\u003cspan\u003e$\u003c/span\u003e 27.5\u0026nbsp;billion) reparation agreement, signed between mining companies and public authorities, was finalized in late 2024, legal battles continue to unfold both domestically and internationally.\u003csup\u003e1\u003c/sup\u003e By the end of 2025, no criminal convictions had been reached in Brazil; meanwhile, in the United Kingdom, a civil action against BHP (listed on the London Stock Exchange at the time) had its first phase concluded in March 2025, but remains pending final judgment (Mansur, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThese enduring repercussions show that, although such events are episodic, their environmental, financial, and legal effects persist over the long term, directly affecting the governance and valuation of the sector\u0026rsquo;s main players. In the disaster-economics literature, this persistence is expected: disasters generate direct and indirect economic losses and can raise perceived risk and risk premia, with consequences that extend beyond the immediate physical shock (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe relevance of this impact is underscored by the strategic position that mining occupies within the Brazilian economy, standing out as one of the most significant sectors in both financial markets and employment generation. In 2024, the sector represented about 4% of Brazil\u0026rsquo;s Gross Domestic Product (GDP), handling approximately 1.24\u0026nbsp;billion tons of minerals and creating over 210,000 direct jobs and an additional 2.2\u0026nbsp;million indirect and induced jobs (IBRAM, 2024). Additionally, the sector was responsible for collecting R\u003cspan\u003e$\u003c/span\u003e 85.6\u0026nbsp;billion (\u0026asymp;\u0026thinsp;US\u003cspan\u003e$\u003c/span\u003e 13.8\u0026nbsp;billion) in taxes and R\u003cspan\u003e$\u003c/span\u003e 6.86\u0026nbsp;billion (\u0026asymp;\u0026thinsp;US\u003cspan\u003e$\u003c/span\u003e 1.1\u0026nbsp;billion) in Financial Compensation for Mineral Exploitation (CFEM), highlighting its economic significance. The leading minerals produced include iron ore, gold, and copper (IBRAM, 2024). Against this backdrop, a natural question arises: how do extreme events (environmental disasters) affect the stock market valuation of firms operating in such a strategic sector?\u003c/p\u003e \u003cp\u003eWithin this setting, Vale S.A., founded in 1942, is the primary representative of the mining sector in Brazil and one of the largest global mining companies. It currently holds a world-leading position in iron ore production, with shares listed on the S\u0026atilde;o Paulo, Paris, Madrid, and New York stock exchanges (Fabr\u0026iacute;cio et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, this economic prominence comes with considerable operational challenges, especially concerning tailings dam safety. Globally, between 1915 and 2019, there were 356 dam failures recorded, reflecting the gravity and frequency of such operational risks (WMTF, 2019). These incidents typically stem from severe natural events, such as earthquakes and storms, or structural and planning failures, including operational and maintenance shortcomings (Duarte, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Beyond the tragic human and environmental toll, the finance-oriented disaster literature stresses that shocks of this nature may also trigger liquidity pressure and heightened risk assessments, with spillovers to corporate financial conditions and market pricing (Halkos \u0026amp; Zisiadou, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn Brazil, particularly in the state of Minas Gerais, the risks associated with tailings dams are frequent and severe. Between 1986 and 2019, at least seven major dam failures were recorded in the state. Among these, the Fund\u0026atilde;o dam rupture in 2015 (controlled by Samarco, a joint venture between Vale and BHP Billiton, in Mariana) stands out as the most catastrophic. This event was Brazil\u0026rsquo;s worst environmental disaster, causing 19 fatalities, widespread community destruction, and severe river contamination (Fernandes et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Such events, often framed as \u0026ldquo;black swans,\u0026rdquo; are rare, difficult to anticipate, and can generate extraordinary impacts that test standard assumptions about risk and information processing in financial markets (Taleb, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Woods, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Masys, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2012\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBrazil, recognized globally as a major commodities exporter with prominent natural resource-based companies, has repeatedly faced serious environmental incidents that affected corporate valuations. Notably, Vale itself experienced another devastating dam collapse in Brumadinho in 2019. Moreover, the oil spills in the Campos Basin involving Chevron in 2011 and Petrobras in 2019 further illustrate how environmental disasters can influence corporate stock prices. In disaster economics, a recurring point is that recovery and post-shock adjustment depend critically on financial channels (access to finance, borrowing, and the broader functioning of credit markets) which helps explain why market-based outcomes may react sharply and, at times, persistently (Keerthiratne \u0026amp; Tol, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In parallel, the literature also discusses risk-mitigation and resilience instruments (e.g., insurance and related financial tools) as part of the policy and governance toolkit (Halkos \u0026amp; Zisiadou, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eConsidering these extreme events, a critical question follows: to what extent can financial markets rapidly price information related to unexpected crises? The Efficient Market Hypothesis (EMH) posits that markets quickly absorb all available information and immediately reflect it in asset prices (Fama, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e1970\u003c/span\u003e; \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1991\u003c/span\u003e). However, rare and high-impact events inherently challenge this view, raising doubts about the market\u0026rsquo;s ability to anticipate and price such occurrences accurately in real time (Taleb, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Woods, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Masys, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2012\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSeveral prior studies have investigated the socioeconomic and environmental impacts of these disasters, including works by Simonato (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), Castro and Almeida (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), Carrillo et al. (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and Biazoli et al. (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Recently, Biazoli et al. (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) advanced this field by applying cluster and synthetic control methods to assess the regional economic impacts of the Mariana disaster. Nevertheless, studies have not specifically examined the isolated effects of the Mariana disaster on Vale S.A.\u0026rsquo;s stock returns, leaving a relevant gap at the intersection of disaster economics and capital markets. This study seeks to fill that gap by evaluating direct financial impacts on stakeholders, particularly through the lens of stock market returns.\u003c/p\u003e \u003cp\u003eAccordingly, the objective of this research is to analyze the impacts of the Mariana disaster on Vale\u0026rsquo;s stock returns using the synthetic control method. This methodology enables the construction of a transparent counterfactual scenario by combining weighted comparable units to estimate how Vale\u0026rsquo;s stock returns would have behaved in the absence of the disaster (Abadie \u0026amp; Gardeazabal, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Abadie et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). This design is especially suitable for settings with a single treated unit and multiple potential controls, because it formalizes the choice of the counterfactual as a data-driven weighted combination rather than relying on a single comparator (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTo strengthen identification and address potential spillovers, we analyze three donor pools: (i) a baseline pool including globally comparable firms indicated in Vale\u0026rsquo;s CVM Reference Form, the mandatory annual disclosure filing for Brazilian listed companies; (ii) a pool excluding Brazilian firms; and (iii) a stricter pool also excluding BHP Group Ltd. The analysis covers 2011 to 2017 and focuses on Mariana as the first major disaster directly associated with Vale within this window, reducing contamination from later shocks.\u003c/p\u003e \u003cp\u003eThis paper contributes in three ways. First, it adds evidence to disaster economics/finance by documenting how a major socio-environmental disaster is priced in equity markets (and for how long) using a flagship commodity producer as a revealing case. Second, it offers a structured synthetic-control design with alternative donor pools and robustness checks (including placebo-based inference), which is particularly useful in rare-event settings. Third, it derives implications for governance, operational risk management, and regulatory oversight in sectors where extreme events can have persistent valuation effects.\u003c/p\u003e \u003cp\u003eOur findings indicate that Vale\u0026rsquo;s stock returns experienced an immediate decline following the disaster, although this initial drop was not as severe as projected by the counterfactual model. Nevertheless, from 2016.1 to 2017.2, Vale\u0026rsquo;s returns remained consistently below the expected values provided by the synthetic control, suggesting that the market initially underestimated the disaster\u0026rsquo;s longer-term impacts. Put differently, while the rupture was a discrete event, its consequences (environmental, financial, social, and legal) continued to shape valuation well beyond the initial shock.\u003c/p\u003e \u003cp\u003eThese results carry scientific and practical implications. Managers may infer the need for stronger investments in operational safety and sustainability, given the potential valuation penalties during and after extreme events. Investors should recognize that valuation approaches anchored in historical returns can understate \u0026ldquo;black swan\u0026rdquo; exposure, strengthening the case for more robust risk-management strategies. Regulators, in turn, may view the evidence as supporting enhanced transparency and preventive measures to strengthen market resilience in the face of rare, high-impact shocks like the Mariana disaster (Rhee \u0026amp; Wu, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e"},{"header":"2. Literature Review","content":"\u003cp\u003eOn November 5, 2015, the Fund\u0026atilde;o tailings dam (operated by Samarco, co-owned by Vale and BHP Billiton) failed in Mariana, Minas Gerais, abruptly releasing an estimated 55\u0026ndash;62 million m\u0026sup3; of iron-ore tailings into the Doce River watershed. The slurry wave engulfed the district of Bento Rodrigues, displacing its residents and causing at least 19 fatalities, while contaminated sediments traveled approximately 663 km along the river system to the Atlantic coast. Beyond immediate destruction, the event posed persistent socioeconomic risks through impacts on water supply, fisheries, agriculture, and other ecosystem services in dozens of municipalities along the Doce River basin (Fernandes et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFrom an economic standpoint, the disaster has also been treated as a long-lived regional shock rather than a short-lived disruption. Using a spatial panel difference-in-differences design, Batista and Firme (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) find that direct economic losses were larger in municipalities closer to the Doce River and decreased with distance, with particularly severe effects in agriculture and industry. Their estimates suggest that direct losses accumulated substantially over the medium run, reaching approximately US\u003cspan\u003e$\u003c/span\u003e 29.1 billion (after 3 years), US\u003cspan\u003e$\u003c/span\u003e 57.1 billion (after 4 years), and US\u003cspan\u003e$\u003c/span\u003e 95.5 billion (after 5 years) (in 2022 prices), while total net effects were smaller due to offsetting spillovers, approximately US\u003cspan\u003e$\u003c/span\u003e 15.7 billion, US\u003cspan\u003e$\u003c/span\u003e 28.0 billion, and US\u003cspan\u003e$\u003c/span\u003e 49.1 billion over the same horizons (Batista \u0026amp; Firme, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003ca class=\"FNLink\" href=\"#Fn2\" id=\"#FNLinkFn2\"\u003e\u003c/a\u003e\u003c/p\u003e \u003cp\u003eThis broader context matters for a disaster-finance perspective: disasters may remain salient for years through litigation, remediation, regulation, and reputational pressure, channels that can reshape expectations about risk and future cash flows even after the \u0026ldquo;news shock\u0026rdquo; has passed.\u003c/p\u003e \u003ch2\u003e2.1. Impacts of Disasters and Efficient Markets\u003c/h2\u003e \u003cp\u003eThe mining sector, given its scale and its potential for severe environmental externalities, carries a non-trivial risk of operational disasters and, as a result, is frequently under public and regulatory scrutiny. As of 2023, Brazil had 927 registered dams, with 464 monitored under the National Dam Safety Policy (PNSB, Portuguese acronym). These structures are classified according to their Risk Category Index (CRI), underscoring the inherent operational risks associated with their management (ANM, 2023). Such risks became evident after the Fund\u0026atilde;o dam rupture in Mariana, widely recognized as one of the world\u0026rsquo;s most severe mining-related environmental disasters.\u003c/p\u003e \u003cp\u003eIncidents of this magnitude expose vulnerabilities in governance, infrastructure integrity, and monitoring systems, reinforcing the need for stronger safety protocols and more transparent environmental reporting (Souza, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Fabr\u0026iacute;cio et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). They also highlight why environmental risks cannot be treated as \u0026ldquo;non-financial\u0026rdquo; in a narrow sense: disaster shocks may reshape expectations about cash flows, liabilities, and risk premia, with valuation consequences that extend beyond the immediate physical impact (Castro \u0026amp; Almeida, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In other words, the disaster-economics literature stresses that disasters generate direct and indirect economic losses and can propagate through financial channels, affecting perceptions of risk and the pricing of affected entities (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis discussion connects naturally to the Efficient Market Hypothesis (EMH). Fama (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e1970\u003c/span\u003e) argues that market prices incorporate available information, leaving little room for persistent abnormal returns. Under this view, new information should be rapidly absorbed by asset prices, making systematic outperformance unlikely. Yet rare and unpredictable events, often labeled \u0026ldquo;black swans\u0026rdquo;, challenge the EMH because they can be difficult to anticipate and may not be immediately priced, especially when uncertainty is high and information arrives in fragments (Taleb, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Woods, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Masys, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Fama (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1991\u003c/span\u003e) further notes that market expectations depend on a broader informational set, implying that assimilation may vary across contexts, particularly during unprecedented crises.\u003c/p\u003e \u003cp\u003eEmpirical evidence also suggests that corporate exposure and recovery are heterogeneous. Chang et al. (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), for instance, show that disaster impacts depend on both internal factors (e.g., firm size, operational flexibility, reliance on local markets) and external conditions (e.g., regulation and government interventions). Insurance coverage and managerial capacity may shape how quickly firms stabilize and how investors revise expectations after shocks (Chang et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). From a public-finance angle, Chen (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) highlights the government\u0026rsquo;s dual burden in crises, shrinking revenues alongside rising expenditures for response and restoration, an institutional setting that can amplify uncertainty around regulation, enforcement, and compensation.\u003c/p\u003e \u003cp\u003eFor economies with large commodity exporters, these mechanisms are particularly salient. Brazil has repeatedly faced environmental incidents that spilled over into corporate valuation debates, including Mariana (in 2015), Brumadinho (in 2019), and major oil spills involving Chevron (in 2011) and Petrobras (in 2019). The point is not merely that \u0026ldquo;prices fall\u0026rdquo; after bad news. Rather, disasters can trigger multi-year processes (remediation, litigation, regulatory revisions, and reputational reassessment) that affect perceived risk and expected cash flows. In that sense, the Mariana case is informative because it illustrates how disaster impacts can persist and diffuse through space and time at the regional level (Batista \u0026amp; Firme, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) and through ecological and socioeconomic channels that may alter livelihoods and local production conditions (Fernandes et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn such settings, the central empirical challenge is not simply documenting price movements, but establishing a credible counterfactual: how would the firm have performed absent the disaster? This is precisely where synthetic control designs are useful. The method constructs a data-driven counterfactual as a weighted combination of comparable units, rather than relying on a single benchmark, and it is especially appropriate when there is one treated unit and many potential controls (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Used carefully, it helps sharpen inference about whether observed post-event dynamics plausibly reflect disaster-related information and risk repricing.\u003c/p\u003e \u003ch2\u003e2.2. Black Swans and Synthetic Control\u003c/h2\u003e \u003cp\u003eThe Mariana disaster illustrates how large corporations can face abrupt market challenges when an extreme event unfolds and information diffuses rapidly. These shocks do not resemble typical macroeconomic fluctuations; they combine operational failure, environmental damage, and evolving legal and political responses. Taleb (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2007\u003c/span\u003e) frames such episodes as \u0026ldquo;black swans\u0026rdquo;: rare, hard-to-predict events with outsized consequences across multiple domains. Long before that label became popular, economic and financial research emphasized that assets exposed to severe downside states may be discounted by risk-averse investors, while assets that remain resilient during crises become relatively more attractive (Roy, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e1952\u003c/span\u003e; Menezes et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e1980\u003c/span\u003e; Rietz, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e1988\u003c/span\u003e; Rhee \u0026amp; Wu, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eEven though EMH and Black Swan Theory start from different premises (price efficiency under rational processing versus the dominance of the improbable), both speak to how economic agents respond to new information flows. The \u0026ldquo;black swan\u0026rdquo; metaphor is useful here mainly because it emphasizes model limits: rare events can invalidate comfortable inferences drawn from long periods of \u0026ldquo;normal times,\u0026rdquo; and crisis narratives may be constructed ex post in ways that appear coherent but do not necessarily reflect ex ante predictability (Taleb, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Krupa \u0026amp; Jones, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). In practical terms, this tendency can lead managers and investors to underestimate tail risks, leaning too heavily on historical regularities that underweight low-probability, high-impact scenarios (Taleb, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Krupa \u0026amp; Jones, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn response to stakeholder pressure and growing awareness of environmental risk, firms in high-impact sectors have increasingly incorporated Corporate Social Responsibility (CSR) and broader sustainability frameworks. The shift toward a \u0026ldquo;triple bottom line\u0026rdquo; perspective, integrating social, environmental, and financial dimensions, also reflects an attempt to address investor and creditor concerns regarding long-run operational continuity and legitimacy (Mart\u0026iacute;nez-Ferrero et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Fabr\u0026iacute;cio et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Still, whether and how quickly markets incorporate these risks remains an empirical question, particularly when disasters trigger prolonged litigation, regulatory uncertainty, and reputational damage.\u003c/p\u003e \u003cp\u003eA growing body of empirical work has examined the Mariana disaster and related events using diverse methods and outcomes. Simonato (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) employs a computable general equilibrium approach and reports persistent negative regional impacts on production, employment, and trade. Castro and Almeida (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) apply synthetic control methods to show significant losses in industrial and mining outputs in affected states after Mariana, highlighting the role of sectoral dependence. Carrillo et al. (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), using a difference-in-differences design, identify adverse health effects among newborns exposed in utero, underscoring that disaster burdens may be unevenly distributed across social groups.\u003c/p\u003e \u003cp\u003eCastro and Almeida (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) extend the synthetic control framework to the Brumadinho disaster, documenting heterogeneous sectoral impacts and recovery patterns across events. More recently, Biazoli et al. (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) combine clustering with synthetic control to build more robust counterfactuals, reinforcing the value of careful comparator construction in disaster settings. Relatedly, Kuruc (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) illustrates synthetic control\u0026rsquo;s usefulness for evaluating crisis-related policy effectiveness, arguing that the approach can yield estimates that are less dependent on standard regression assumptions.\u003c/p\u003e \u003cp\u003eBatista and Firme (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) complement this evidence by explicitly modeling spatial spillovers and heterogeneous exposure across \u0026ldquo;neighbourhood levels\u0026rdquo;, documenting that losses were strongest closer to the Doce River and smaller farther away, while remaining economically meaningful in the medium run. Their results are helpful for a capital-markets lens because they underscore that Mariana was not a purely local and short-lived shock; rather, it generated a persistent and spatially structured economic footprint, a feature that can sustain investor attention and uncertainty over time (Batista \u0026amp; Firme, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWhile much of this literature focuses on regional output, health, or sectoral activity, less attention has been given to isolating the disaster\u0026rsquo;s effect on the stock-market performance of a directly involved firm. This gap matters because market prices summarize expectations about cash flows, risk, and uncertainty, and they are a key channel through which disasters affect firms\u0026rsquo; cost of capital and governance pressures. In this study, we therefore focus on Vale\u0026rsquo;s stock returns and use synthetic control methods to construct an explicit counterfactual path for returns in the absence of the Mariana shock. Given this framing, we articulate the following testable expectations:\u003cp\u003eH1: Vale\u0026rsquo;s stock returns declined relative to the synthetic-control counterfactual immediately after the Mariana disaster.\u003c/p\u003e\u003cp\u003eH2: Vale\u0026rsquo;s stock returns remained below the synthetic-control counterfactual in the post-disaster period, consistent with delayed market adjustment to the disaster\u0026rsquo;s longer-run implications.\u003c/p\u003e\u003c/p\u003e \u003cp\u003eFrom these hypotheses, we develop an empirical strategy based on synthetic control methods, using comparable firms to build a robust counterfactual scenario. This design choice is aligned with recent contributions in the disaster-economics and disaster-finance literature that rely on transparent counterfactual construction when a single unit is exposed to a major shock (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003ch2\u003e2.3. Disaster-Economics and disaster-finance lens: channels to market pricing\u003c/h2\u003e \u003cp\u003eTo connect the disaster-economics core of the journal to a capital-markets setting, it is useful to be explicit about the channels through which disasters can affect market pricing. First, disasters often generate large losses and sudden liquidity needs, while simultaneously increasing uncertainty about operational continuity and future liabilities. In settings where access to finance is limited or uneven, these frictions can shape recovery dynamics and the way losses translate into economic outcomes (Keerthiratne \u0026amp; Tol, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). This also raises an identification concern: a recurring issue in empirical disaster research is that common disaster measures can embed human and institutional components, complicating causal interpretation and strengthening the case for designs that emphasize credible counterfactuals and careful identification (Keerthiratne \u0026amp; Tol, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSecond, disaster-risk management is frequently framed as a component of resilience. Financial instruments (such as insurance, hedging strategies, and catastrophe-linked securities) are discussed in the literature as mechanisms that can transfer risk and smooth post-disaster fiscal or corporate stress (Halkos \u0026amp; Zisiadou, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In this sense, market reactions to disasters can be interpreted as rapid updates in expectations about risk, financing conditions, and governance quality, not merely as responses to direct physical damage.\u003c/p\u003e \u003cp\u003eFinally, these channels motivate why synthetic control is a useful design for disaster-finance questions. By producing an explicit counterfactual that does not depend on a single comparison unit, the method helps clarify whether observed valuation dynamics plausibly reflect disaster-related repricing rather than idiosyncratic contemporaneous shocks. Recent work has used synthetic control designs to evaluate disaster-related policy and financial mechanisms, precisely because the appeal lies in constructing a transparent counterfactual as a weighted combination of controls (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This framing keeps the Vale case as an empirical trigger, while positioning the broader contribution as evidence on how environmental disasters may be (mis)priced (and for how long) in capital markets.\u003c/p\u003e "},{"header":"3. Method","content":"\u003cp\u003eThis section describes the empirical strategy used to evaluate how the Fund\u0026atilde;o tailings dam collapse (Mariana, Brazil; November 5, 2015) affected Vale S.A.\u0026rsquo;s stock returns, and whether the observed price dynamics are consistent with rapid versus delayed adjustment in the presence of a rare, high-impact shock. The treated unit is Vale, the outcome is the semiannual cumulative stock return, and the treatment occurs in 2015.2 (second semester of 2015). The analysis window spans 2011.2-2017.2, allowing us to compare Vale\u0026rsquo;s realized returns to a transparent counterfactual path built from comparable firms. Consistent with the hypotheses developed in the literature review, we interpret the estimated post-event gaps in two layers: (i) the immediate deviation in the disaster semester (H1), and (ii) the persistence of deviations in subsequent semesters (H2).\u003c/p\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Construction of Synthetic Control for Vale\u003c/h2\u003e \u003cp\u003eTo evaluate the impact of the Mariana disaster on the stock returns of Vale, it is essential to compare the company\u0026rsquo;s actual performance after the environmental incident with a hypothetical scenario in which the disaster did not occur. This hypothetical or counterfactual scenario cannot be directly observed. Thus, constructing a robust and reliable counterfactual becomes crucial for performing such comparisons.\u003c/p\u003e \u003cp\u003eIn this study, we employ the synthetic control method (SCM), originally proposed by Abadie and Gardeazabal (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2003\u003c/span\u003e) and further developed by Abadie et al. (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). This technique is particularly suited for comparative case studies using aggregate data or situations involving few treated units, such as the specific case of Vale analyzed here. Unlike traditional methods like difference-in-differences, SCM emphasizes strong pre-treatment fit and an explicit counterfactual built from a weighted combination of controls, which reduces reliance on strong functional-form assumptions and helps clarify identification in settings where confounding and measurement challenges are common (Rosenbaum, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Keerthiratne \u0026amp; Tol, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). A key step in this design is defining an appropriate donor pool and a predictor set so that the treated unit and the synthetic unit are as comparable as possible before the shock (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe synthetic control method constructs a weighted average of selected control units (other companies with similar characteristics to Vale) to replicate the counterfactual trajectory of Vale\u0026rsquo;s stock returns. The weights attributed to these control units aim to minimize the Root Mean Squared Prediction Error (RMSPE) between the actual and synthetic trajectories before the disaster (Abadie et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Consequently, the synthetic trajectory closely replicates the real trajectory observed before the event, enabling a meaningful analysis of the disaster\u0026rsquo;s impact. In operational terms, the estimated \u0026ldquo;effect\u0026rdquo; for each semester \u003cem\u003et\u003c/em\u003e is the gap between Vale\u0026rsquo;s observed return and the synthetic return:\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:{\\alpha\\:}_{t}={SR}_{t}^{Vale}-{SR}_{t}^{Synth}\\)\u003c/span\u003e \u003c/span\u003e​\u003c/p\u003e \u003cp\u003eWhere, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{SR}_{t}^{Synth}\\)\u003c/span\u003e\u003c/span\u003e is the weighted sum of donor-firm returns using the SCM weights.\u003c/p\u003e \u003cp\u003eHowever, an essential assumption of the synthetic control method is the absence of spillover effects, where the event affects control units indirectly, potentially contaminating the counterfactual (Rosenbaum, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). In our analysis, spillover effects could manifest through impacts on iron ore prices, given Vale\u0026rsquo;s prominent global role. As Castro and Almeida (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) highlight, disasters involving major commodity producers like Vale may significantly influence global commodity prices, thus affecting comparable companies and complicating the construction of the synthetic trajectory.\u003c/p\u003e \u003cp\u003eIn this setting, two macro-financial variables are particularly relevant: international iron ore prices and the U.S. dollar index, because they may affect mining firms\u0026rsquo; returns. However, because these variables are common time shocks (i.e., they do not vary across firms in a given semester), they do not provide the cross-sectional variation required for the SCM optimization stage. For this reason, they were not included as predictors in the synthetic-control optimization, even though they remain economically meaningful drivers of returns in commodity-linked industries (Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). In practice, the SCM design addresses such common shocks through pre-treatment matching and by comparing Vale to a counterfactual formed from firms exposed to the same global conditions.\u003c/p\u003e \u003cp\u003eVale operates globally, holding significant production, sales, and distribution activities across various countries. As illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Vale is among the world\u0026rsquo;s leading iron ore producers, with a strong international market presence. In Brazil\u0026rsquo;s stock market, Vale is classified under Basic Materials in the \u0026ldquo;Metallic Minerals\u0026rdquo; segment. Alongside Vale, this segment includes only seven other companies: Aura Almas Minera\u0026ccedil;\u0026atilde;o S.A., Aura Minerals Inc., Bradespar S.A., Companhia Brasileira de Alum\u0026iacute;nio, CSN Minera\u0026ccedil;\u0026atilde;o S.A., Litel Participa\u0026ccedil;\u0026otilde;es S.A., and Litela Participa\u0026ccedil;\u0026otilde;es S.A.\u003c/p\u003e\u003cp\u003eGiven Vale\u0026rsquo;s significance in the global market and potential spillover risks, special attention was paid to selecting appropriate donor pools (control groups) to mitigate such risks. Following Castro and Almeida\u0026rsquo;s (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) logic of mitigating contamination, while noting that their focus is regional economic outcomes and ours is firm-level stock returns, we adopted distinct donor pools to avoid contamination of the control units. The donor pool selection followed a two-step strategy. Initially, all competitor companies mentioned in Vale\u0026rsquo;s Reference Form were considered. Subsequently, due to data availability constraints, only companies present in the LSEG database were included, as detailed in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eQuantity and selection criteria of Vale\u0026rsquo;s donor pools\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eQuantity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRationale (brief)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompanies cited in Vale\u0026rsquo;s Reference Form\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eInitial universe of peers disclosed by Vale\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(-) Companies not available in LSEG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(11)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eData availability constraint\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e= Donor pool 1 (DP1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBaseline eligible donor pool\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(-) Brazilian companies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReduce potential domestic spillovers\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e= Donor pool 2 (DP2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStricter donor pool (international only)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(-) BHP Group Ltd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAvoid direct involvement with Samarco\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e= Donor pool 3 (DP3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStrictest donor pool\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eSource: Research data (2026).\u003c/p\u003e \u003cp\u003eThe construction of three distinct donor pools was specifically designed to address potential spillover concerns systematically. DP1 comprises all identified companies initially available. DP2 excludes Brazilian firms (CSN, Gerdau, and Usiminas), as they might have experienced indirect spillover effects due to the disaster\u0026rsquo;s impact on the national market. DP3 further excludes BHP Group Ltd., a co-owner of Samarco (operator of the Fund\u0026atilde;o dam), to avoid potential direct contamination from spillover effects.\u003c/p\u003e \u003cp\u003eThe predictor variables utilized to construct the synthetic control were selected based on solid theoretical foundations and previous empirical studies, including Debt and Net Income (Sampaio, Azevedo \u0026amp; Azuaga, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Additionally, to enhance the robustness of the synthetic control, firm size (market capitalization) and ESG controversies were included.\u003c/p\u003e \u003cp\u003eThe ESG Controversies Score is a scale of 0 to 100, measuring the company\u0026rsquo;s exposure to controversies/scandals and/or violations of ESG pillars captured by the media. A higher score indicates lower exposure to scandals. Therefore, this proxy provides an external perspective of the company (LSEG, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). The literature demonstrates a relationship between extreme environmental events and ESG disclosures and measures (Huang, Li, Lin, \u0026amp; McBrayer, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The financial market itself ends up reacting negatively to companies that experience some type of scandal or suspicious activities, which impacts their market value (Aouadi \u0026amp; Marsat, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Dalcero et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBecause SCM estimates can be sensitive to donor-pool composition and predictor choices, we report the three donor-pool constructions transparently and use them as structured sensitivity checks. This is consistent with methodological discussions emphasizing that robustness should be evaluated by examining whether estimated post-event gaps are driven by specific donor units or by the broader quality of the pre-treatment fit (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTherefore, the meticulous construction of donor pools provides methodological robustness, enabling the analysis of Vale\u0026rsquo;s stock price responses to the Mariana disaster. This careful construction enhances the validity of the study, allowing us to evaluate the impact of rare and highly impactful incidents (Black Swan events) on corporate financial markets.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.2. Data Collection and Analysis\u003c/h2\u003e \u003cp\u003eThe observational units adopted in this study were Vale\u0026rsquo;s national and international competitor firms. We used stock returns as a proxy for companies\u0026rsquo; economic and financial performance, allowing us to analyze how the financial market responded to the Mariana disaster. Given that the study included 16 companies from various countries: Australia (BHP Group Ltd, Fortescue Ltd, Rio Tinto Ltd), Brazil (Companhia Sider\u0026uacute;rgica Nacional SA, Gerdau SA, Usinas Sider\u0026uacute;rgicas de Minas Gerais SA), Canada (First Quantum Minerals Ltd), China (Jiangxi Copper Co Ltd), France (Eramet SA), Germany (Aurubis AG), South Africa (Kumba Iron Ore Ltd), Switzerland (Ferrexpo PLC, Glencore PLC), United Kingdom (Anglo American PLC, Antofagasta PLC), and the United States (Freeport-McMoRan Inc).\u003c/p\u003e \u003cp\u003eWe needed to adjust accounting data to accommodate the shortest available reporting interval in the database. Hence, stock returns were adjusted to a semiannual basis, considering that international companies in our sample report their financial statements semiannually in the LSEG database. Throughout the paper, \u0026ldquo;.1\u0026rdquo; denotes the first semester (January-June) and \u0026ldquo;.2\u0026rdquo; denotes the second semester (July-December).\u003c/p\u003e \u003cp\u003ePrevious studies that examined the environmental disaster\u0026rsquo;s impact on Vale\u0026rsquo;s returns primarily adopted the event study methodology, focusing on cumulative returns around the event date (Foga\u0026ccedil;a, Raeder \u0026amp; Marques, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Rocha \u0026amp; Vasconcelos, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Although our research employs a different approach, it is important to acknowledge that the broader temporal aggregation we use can limit precision in capturing very short-lived immediate effects. However, this broader scope is well aligned with our objective of investigating whether disaster-related repricing persists beyond the initial shock, which is central to testing delayed adjustment in the semesters following the event.\u003c/p\u003e \u003cp\u003eMonthly returns for the analysis period were directly extracted from the LSEG Eikon Datastream. Once obtained, we calculated cumulative returns for each semester. The use of stock returns as the main variable is grounded in the financial literature based on the Efficient Market Hypothesis (EMH), which posits that financial asset prices reflect all available information, both public and private (Fama, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e1970\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1991\u003c/span\u003e). Additionally, this variable is widely employed in research examining the financial market impacts of adverse events (Foga\u0026ccedil;a, Raeder \u0026amp; Marques, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Rocha \u0026amp; Vasconcelos, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Additional information about the variables used can be found in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDescription of variables used\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAbbreviation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eType\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSource\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStock Return\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDependent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSemiannual cumulative return =\u003c/p\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\prod\\:}_{m=1}^{6}\\left(1+{RM}_{m}\\right)\\)\u003c/span\u003e\u003c/span\u003e-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLSEG Eikon\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFirm Size\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSIZE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eProxy used was Company Market Capitalization\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLSEG Eikon\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDebt\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDEB\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(Short-term and long-term interest-bearing liabilities) / Total Assets\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLSEG Eikon\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eESG Controversies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eESG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMethodology of the database itself\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLSEG Eikon\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNet Income\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eProxy used was Net Income\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLSEG Eikon\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eSource: Research data (2026).\u003c/p\u003e \u003cp\u003eThe predictor variables derived from company financial statements were aggregated semiannually, which corresponds to the smallest available time frame for the entire sample. The analysis covers the period from the second half of 2011 to the second half of 2017, comprising 13 semesters. Given that the Fund\u0026atilde;o dam collapse in Mariana occurred on November 5, 2015, the semesters studied were divided as follows: 8 semesters pre-disaster (2011.2 to 2015.1), 1 semester during the disaster (2015.2), and 4 semesters post-disaster (2016.1 to 2017.2).\u003c/p\u003e \u003cp\u003eThe starting period was chosen due to data availability for all analyzed companies (e.g., Glencore PLC started trading in the first semester of 2011, making its data available from 2011.2 onwards). The endpoint of our analysis was a discretionary choice that does not compromise our findings. According to Abadie et al. (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), the synthetic control technique yields optimal results when the number of observations before the event is larger than the number of observations after it.\u003c/p\u003e \u003cp\u003eFinally, to connect the empirical design to the hypotheses, we interpret the post-event gaps \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\alpha\\:}_{t}\\)\u003c/span\u003e\u003c/span\u003e as follows. H1 is assessed by examining the sign and magnitude of the gap in 2015.2 (the disaster semester). H2 is assessed by examining whether the gaps remain negative in the post-disaster semesters (2016.1-2017.2), consistent with persistent underperformance relative to the counterfactual. In line with standard SCM practice, inference is supported by transparency in pre-treatment fit (RMSPE) and by placebo-style comparisons that evaluate whether the estimated post-event path for Vale is unusual relative to gaps obtained when the \u0026ldquo;treatment\u0026rdquo; is reassigned to donor units (Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). These procedures complement the structured sensitivity analysis provided by DP1-DP3 and help assess whether the estimated disaster-related deviation is plausibly distinct from the distribution of placebo gaps.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Results Analysis","content":"\u003cp\u003eThis section reports and interprets the synthetic-control estimates for Vale using three donor pools: DP1 (all firms in the final sample), DP2 (excluding Brazilian firms to mitigate potential domestic spillovers), and DP3 (excluding both Brazilian firms and BHP Group Ltd., given its direct link to Samarco at the time of the disaster). The goal is to test our two expectations from the literature: (H1) an immediate negative gap around the disaster semester (2015.2), and (H2) a persistent post-disaster gap consistent with delayed market adjustment. Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e summarizes the donor-pool composition and the resulting pre-disaster fit (RMSPE) for each specification.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCompanies\u0026rsquo; contribution to weight formation in DP1, DP2, and DP3\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompanies\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCountry\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTicker\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDP1\u003c/p\u003e \u003cp\u003eWeight\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDP2\u003c/p\u003e \u003cp\u003eWeight\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eDP3\u003c/p\u003e \u003cp\u003eWeight\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnglo American PLC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnited Kingdom\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAAL.L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.049\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.064\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAntofagasta PLC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnited Kingdom\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eANTO.L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.028\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.039\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAurubis AG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGermany\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNAFG.DE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.033\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.044\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBHP Group Ltd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAustralia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBHP.AX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.315\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.315\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eX\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompanhia Sider\u0026uacute;rgica Nacional SA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBrazil\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCSNA3.SA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.009\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eX\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEramet SA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFrance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eERMT.PA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.036\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.047\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFerrexpo PLC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSwitzerland\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFXPO.L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFirst Quantum Minerals Ltd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCanada\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFM.TO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.040\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFortescue Ltd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAustralia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFMG.AX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.064\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFreeport-McMoRan Inc\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnited States\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFCX.N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.031\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.046\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e-\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGerdau SA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBrazil\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGGBR4.SA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eX\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlencore PLC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSwitzerland\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGLEN.L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.121\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.109\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eJiangxi Copper Co Ltd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e600362.SS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.076\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKumba Iron Ore Ltd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSouth Africa\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eKIOJ.J\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.035\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.045\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.147\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRio Tinto Ltd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAustralia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRIO.L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.159\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.158\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.789\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUsinas Siderurgicas de Minas Gerais SA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBrazil\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUSIM5.SA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eX\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRMSPE\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.00916\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.00890\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.00561\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNote: RMSPE is the Root Mean Squared Prediction Error; \u0026ldquo;X\u0026rdquo; means the company was excluded from the donor pool; \u0026ldquo;-\u0026rdquo; indicates the company was available but not selected for synthetic control. The sum of the weights in each column equals 1 (100%). Source: Research data (2026).\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e indicates that DP1 provides the best pre-disaster fit, with the lowest RMSPE (0.00916). DP1 also yields a broadly distributed synthetic composition, with meaningful contributions from several global peers. In DP1, the largest weights are assigned to BHP Group (0.315), Rio Tinto (0.159), and Glencore (0.121), followed by Jiangxi Copper (0.076). This pattern is consistent with the logic of SCM: the synthetic unit is constructed as a weighted combination of controls that best matches the treated unit in the pre-event period (Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Importantly, this is also the reason donor-pool definition and pre-treatment comparability are central design choices in SCM applications (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDP2, which excludes Brazilian firms to reduce potential national-market spillovers, delivers a weaker (though still acceptable) pre-disaster fit (RMSPE 0.00890). The weighting structure remains diversified, with sizable contributions from BHP (0.315), Rio Tinto (0.158), and Glencore (0.109), plus moderate weights spread across other international peers. In substantive terms, DP2 functions as a spillover-robustness check: it reduces the likelihood that domestic firms\u0026rsquo; co-movement with Vale (driven by shared Brazilian macro-financial conditions) mechanically improves fit.\u003c/p\u003e \u003cp\u003eDP3 is the most restrictive specification, excluding Brazilian firms and BHP Group Ltd. Given BHP\u0026rsquo;s direct association with Samarco, this exclusion is conceptually appealing from a contamination standpoint, but it comes with a clear trade-off: DP3 delivers the poorest pre-disaster fit (RMSPE 0.00561) and produces a highly concentrated synthetic control. Only three firms receive positive weights, dominated by Rio Tinto (0.789), followed by Kumba Iron Ore (0.147) and Fortescue (0.064). This concentration heightens sensitivity to idiosyncratic features of a single comparator, an issue often discussed as a practical limitation of restrictive donor pools in SCM designs (Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). For that reason, we treat DP3 primarily as a robustness bound, rather than as the preferred baseline.\u003c/p\u003e \u003cp\u003eTo complement Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e summarizes (i) predictor balance between Vale and its synthetic counterpart (illustrative for DP1), and (ii) the relative importance (\u0026ldquo;predictor weights\u0026rdquo;) assigned during optimization. For readability and replication, monetary values are reported in US\u003cspan\u003e$\u003c/span\u003e billions (rounded), while ratios and scores retain their natural scales.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eContribution of variables to weight formation in DP1, DP2, and DP3\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTreated\u003c/p\u003e \u003cp\u003e(Vale)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSynthetic\u003c/p\u003e \u003cp\u003e(DP1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSample\u003c/p\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDP1\u003c/p\u003e \u003cp\u003eWeight\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eDP2\u003c/p\u003e \u003cp\u003eWeight\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDP3\u003c/p\u003e \u003cp\u003eWeight\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFirm Size (Market Cap, US\u003cspan\u003e$\u003c/span\u003e bn)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e76.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e76.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e28.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.130\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.540\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDebt (Debt/Total Assets)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.229\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.229\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.261\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.405\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.479\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.458\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNet Income (US\u003cspan\u003e$\u003c/span\u003e bn)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.900\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.900\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.520\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.093\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.027\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eESG Controversies Score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e50.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e77.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.372\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.493\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRMSPE\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.00916\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.00890\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.00561\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eNote: \u0026ldquo;-\u0026rdquo; indicates the predictor was available but not selected (i.e., received zero weight) in that donor-pool optimization. The sum of the weights in each column equals 1 (100%). Source: Research data (2026).\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAcross specifications, Debt receives the largest (or near-largest) predictor weight (DP1\u0026thinsp;=\u0026thinsp;0.405; DP2\u0026thinsp;=\u0026thinsp;0.479; DP3\u0026thinsp;=\u0026thinsp;0.458), suggesting that capital-structure similarity is pivotal for reproducing Vale\u0026rsquo;s pre-disaster return trajectory. This is economically coherent in a disaster-finance setting: leverage is a salient risk indicator when shocks raise uncertainty about liabilities, cash flows, and financing conditions. In parallel, the ESG Controversies Score receives substantial weight in DP1 (0.372) and DP2 (0.493), consistent with the view that scandals/controversies can be material for valuation and can interact with how markets process disaster-related news (Huang et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Aouadi \u0026amp; Marsat, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Dalcero et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The prominence of these predictors supports a disaster-finance interpretation in which the post-event gap may reflect repricing of risk, financing conditions, and perceived governance quality after the shock (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Halkos \u0026amp; Zisiadou, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHaving established donor-pool fit and predictor structure, we now turn to the estimated effects in the event semester and post-event period. Figures\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e plot Vale\u0026rsquo;s observed returns against the synthetic counterfactual for each donor pool. For consistency with the two hypotheses, we discuss results in two steps: (i) the disaster semester (2015.2), aligned with H1, and (ii) the post-disaster semesters (2016.1-2017.2), aligned with H2.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSource: Prepared by the authors (2026).\u003c/p\u003e \u003cp\u003eIn DP1, the pre-disaster period (2011.2-2015.1) shows a close alignment between Vale and its synthetic counterpart, consistent with the extremely low RMSPE reported in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Minor deviations are expected even under strong fit and do not undermine identification, as SCM inference hinges on the overall quality of the pre-event match (Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cb\u003eTesting H1 (immediate effect in 2015.2)\u003c/b\u003e: in the disaster semester (2015.2), Vale\u0026rsquo;s observed returns decline, indicating an immediate negative response aligned with the idea that severe shocks can be capitalized into prices. Interestingly, the synthetic series shows an even sharper decline in the same semester. Rather than treating this as a contradiction, it is more informative to interpret it as evidence that immediate pricing may be shaped by offsetting forces: while disaster news increases uncertainty and perceived tail risk, short-run price dynamics may also embed expectations about remediation, commodity-market conditions, firm prominence, and investors\u0026rsquo; prior beliefs.\u003c/p\u003e \u003cp\u003eFrom a disaster-finance lens, what matters is that disasters can trigger abrupt updates in perceived downside risk and liquidity/financing constraints, channels that can appear with lags or be partially masked in the immediate window (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Thus, H1 is supported in terms of sign (returns fall), but the DP1 comparison suggests that the immediate \u0026ldquo;gap\u0026rdquo; is not the main margin of adjustment in this case.\u003c/p\u003e \u003cp\u003e \u003cb\u003eTesting H2 (persistence after 2015.2)\u003c/b\u003e: the more pronounced and economically meaningful result emerges in the post-disaster period. From 2016.1 onward, Vale\u0026rsquo;s returns remain systematically below the synthetic counterfactual for an extended sequence of semesters. This pattern is consistent with delayed adjustment: markets may underreact initially when the long-run scope of legal, regulatory, and reputational consequences is difficult to quantify, and then gradually incorporate disaster-related information as uncertainties resolve.\u003c/p\u003e \u003cp\u003eThis is precisely the type of mechanism emphasized in disaster-economics/finance research: beyond immediate physical losses, disasters can propagate through financial channels and reshape risk premia and financing conditions over time (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In the Mariana context, prolonged litigation and compensation uncertainty are plausible channels sustaining risk salience, which can translate into persistent valuation discounts.\u003c/p\u003e \u003cp\u003eAlthough DP1 is the baseline due to superior pre-event fit, DP2 and DP3 provide useful robustness checks. In DP2, the pre-event tracking remains visually close (though weaker than DP1 in RMSPE terms), and the post-event pattern continues to suggest that Vale underperforms its counterfactual path over multiple semesters. In DP3, the comparison should be interpreted with greater caution, given its high RMSPE and heavy reliance on a single firm (Rio Tinto). Even so, DP3 is directionally consistent with the broader narrative: the post-event period exhibits a sustained divergence that is difficult to reconcile with instantaneous and complete adjustment. Taken together, these comparisons strengthen the inference that the main empirical phenomenon is persistence, not merely an immediate dip.\u003c/p\u003e \u003cp\u003eFigures \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e present robustness analyses with DP2 and DP3.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSource: Prepared by the authors (2026).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSource: Prepared by the authors (2026).\u003c/p\u003e \u003cp\u003eTo assess whether the estimated post-event divergence could be driven by idiosyncratic noise rather than the disaster shock, we performed a placebo test in which each donor-pool firm (in DP1) is iteratively treated as if it experienced the Mariana disaster. This produces a distribution of placebo gaps that serves as an informal benchmark for the \u0026ldquo;unusualness\u0026rdquo; of Vale\u0026rsquo;s estimated effect (Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Figure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e summarizes these placebo trajectories.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSource: Prepared by the authors (2026).\u003c/p\u003e \u003cp\u003eBefore 2015.2, placebo trajectories tend to fluctuate around zero with limited dispersion, which is expected if pre-event fit is broadly adequate across units. After 2015.2, dispersion widens, reflecting heterogeneity in firms\u0026rsquo; return dynamics and exposure to global commodity-cycle movements and firm-specific shocks. The key interpretive step is not simply that \u0026ldquo;dispersion exists,\u0026rdquo; but whether Vale\u0026rsquo;s post-event gap is large relative to the placebo distribution, especially when benchmarked against pre-event fit. In SCM practice, this logic is often strengthened by comparing post-to-pre RMSPE ratios (or related normalized metrics), which helps distinguish large effects from artifacts of poor fit (Abadie et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). In our setting (where DP1 fit is extremely strong), the sustained negative divergence for Vale in the post-disaster period is consistent with a meaningful disaster-related repricing rather than random fluctuation.\u003c/p\u003e \u003cp\u003eTwo implications follow from the pattern of results. First, the results suggest that the Mariana disaster was not fully priced as a one-off \u0026ldquo;news shock\u0026rdquo;. Instead, pricing appears to incorporate the disaster\u0026rsquo;s consequences more gradually, consistent with the idea that disasters can create persistent financial frictions, through uncertainty about liabilities, governance and compliance expectations, and the perceived reliability of risk controls (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSecond, the evidence speaks directly to governance and resilience: if markets penalize firms for prolonged periods after major socio-environmental disasters, then investments in disaster-risk management, preparedness, and transparency can be interpreted not only as ethical or regulatory priorities but also as strategies to reduce valuation losses under tail risks (Halkos \u0026amp; Zisiadou, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In capital markets, where prices aggregate beliefs about risk and future cash flows, these mechanisms can translate into sustained discounts, particularly in high-impact sectors where operational failures are tightly linked to regulatory scrutiny and stakeholder pressure.\u003c/p\u003e \u003cp\u003eOverall, the results align with our framing in the Introduction and Literature Review: H1 is supported in the sense that returns decline in the disaster semester, but the central evidence is the persistent post-event gap consistent with H2, suggesting delayed market adjustment following a rare, high-impact socio-environmental shock.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThe evidence from this study points to a clear takeaway: in a major socio-environmental disaster, the most consequential valuation effects may arise not only from the initial shock but from the persistence of uncertainty and risk repricing over subsequent periods. For Mariana, the results indicate a short-run drop in Vale\u0026rsquo;s returns in the disaster semester (2015.2), followed by a more informative pattern: a sustained underperformance relative to a credible counterfactual in 2016.1-2017.2. This dynamic is difficult to reconcile with strict interpretations of instantaneous, full incorporation of disaster-related information, and instead aligns with delayed price adjustment under a rare, high-impact event.\u003c/p\u003e \u003cp\u003eThe immediate semester of the rupture provides an important nuance. Vale\u0026rsquo;s return decline in 2015.2 is consistent with an adverse market reaction (H1), yet the drop is not as severe as the synthetic benchmark would suggest. Rather than treating this as a paradox, a more plausible reading is that early pricing combined multiple, partially offsetting signals: the shock itself, expectations about compensation and remediation, and uncertainty about the timing and scale of liabilities. In disaster contexts, information rarely arrives as a single \u0026ldquo;clean\u0026rdquo; announcement (legal developments, enforcement, operational constraints, and reputational damage often unfold in fragments). Under such conditions, initial returns can reflect a provisional narrative that is later revised.\u003c/p\u003e \u003cp\u003eThe post-disaster trajectory is therefore the core result. The persistent gap between Vale and the synthetic counterfactual (H2) suggests that the market initially underweighted longer-run consequences and gradually incorporated them as uncertainty evolved. From a disaster-finance lens, this pattern is economically meaningful because it is consistent with repricing of perceived downside risk, liquidity pressure, and financing conditions, channels through which disasters can affect valuation beyond direct physical losses (Maran, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In practical terms, the market appears to have treated Mariana not merely as a one-off accident but as the beginning of a multi-period process with governance, legal, and operational implications.\u003c/p\u003e \u003cp\u003eThis paper contributes to the literature in three ways, and each contribution is anchored in what the results can (and cannot) support. First, it offers firm-level evidence on how an extreme socio-environmental disaster is (mis)priced over time in equity markets. Much of the existing Mariana-related evidence emphasizes regional economic activity, sectoral outcomes, or health impacts; our focus complements those perspectives by isolating a stakeholder-relevant capital-markets outcome and tracing its persistence. Second, the paper provides a structured synthetic-control design that makes donor-pool sensitivity visible rather than implicit. By contrasting DP1, DP2, and DP3, the study shows how spillover concerns and comparability trade-offs can change pre-fit and inference, which is precisely the type of transparency expected in applied disaster-economics work. Third, the placebo-style exercise supports the interpretation that the post-event divergence is not merely noise, reinforcing the credibility of the counterfactual-based narrative when pre-event fit is strong.\u003c/p\u003e \u003cp\u003eThe implications follow directly from the persistence result. For investors, the evidence cautions against reading disaster risk primarily through short event windows or \u0026ldquo;first-day\u0026rdquo; market reactions. When disasters trigger prolonged litigation, regulatory adjustments, and reputational reassessment, early price movements may understate the longer-run valuation penalty. This matters for portfolio construction, for scenario analysis, and for any valuation approach that extrapolates from historical returns while underweighting tail risk.\u003c/p\u003e \u003cp\u003eFor managers, especially in environmentally sensitive sectors, our findings underscore that disaster-risk management is not a peripheral compliance item. It can be interpreted as a value-preserving component of resilience that shapes how shocks translate into perceived risk and expected cash flows. Are operational safety systems, disclosure practices, and governance arrangements designed to remain credible under stress, or only under normal times? The broader disaster-finance literature highlights that preparedness and financial mechanisms (including insurance and related instruments) may influence how shocks propagate through corporate financial stress and market expectations, reinforcing the relevance of integrating disaster-risk thinking into governance and financial planning (Halkos \u0026amp; Zisiadou, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor regulators and policymakers, the results support a concrete rationale for stronger oversight and more credible disclosure regimes around operational and socio-environmental risks. If markets may underreact initially and adjust gradually as uncertainty resolves, then improving transparency and enforcement can reduce informational frictions that prolong valuation discounts and weaken trust. In tailings dam contexts, this points to the importance of monitoring systems, clear liability frameworks, and disclosure standards that allow stakeholders to assess both near-term exposure and longer-run contingent risks.\u003c/p\u003e \u003cp\u003eThis study has limitations that also define a feasible agenda for future research. The sample window is designed to preserve a strong pre-treatment fit and minimize contamination from later shocks, but it may not capture effects that extend further into the long run. In addition, the predictor set is constrained by database coverage and may not fully reflect all dimensions relevant for return dynamics (e.g., richer governance proxies or alternative risk measures). Future work can extend the horizon, explore complementary outcomes (volatility, downside risk, cost-of-capital proxies), and compare Mariana with other disasters, to test whether delayed adjustment is a recurring feature of socio-environmental shocks or varies with institutional response, media salience, and litigation dynamics.\u003c/p\u003e \u003cp\u003eIn sum, the main message is specific: the Mariana disaster is associated with an immediate negative return shock, but the more informative evidence lies in the persistent post-disaster gap relative to a credible counterfactual, consistent with gradual repricing of risk and uncertainty rather than instantaneous incorporation. This is the \u0026ldquo;so what\u0026rdquo; for disaster finance: governance quality, transparency, and risk management can plausibly influence not only the size of the initial valuation impact, but also how long the market continues to discount the firm after an extreme event.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eCompeting interests\u003c/h2\u003e \u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThe authors received no specific funding for this work.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eMatheus Soares Mendes, Jo\u0026atilde;o Antonio Da Costa Neto, Francisco De Assis Miranda Da Silva, and Orleans Silva Martins jointly conceived and designed the research project. All authors contributed equally to the development of the research question, theoretical framing, and formulation of hypotheses. They collaboratively conducted the literature review, designed the empirical strategy, and defined the methodological procedures. Data collection, organization, and analysis were performed in a coordinated manner, with all authors actively participating in the interpretation of results and discussion of implications. The manuscript was written collectively, with iterative rounds of drafting and revision involving all authors. All authors critically reviewed the final version, approved it for submission, and agree to be accountable for all aspects of the work.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data that support the findings of this study are available from third-party providers (e.g., LSEG) under license restrictions. Replication code and derived data used in the analysis can be made available by the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAbadie A, Diamond A, Hainmueller J (2010) Synthetic control methods for comparative case studies: estimating the effect of California\u0026rsquo;s tobacco control program. J Am Stat Assoc 105(490):493\u0026ndash;505. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1198/jasa.2009.ap08746\u003c/span\u003e\u003cspan address=\"10.1198/jasa.2009.ap08746\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbadie A, Diamond A, Hainmueller J (2015) Comparative politics and the synthetic control method. Am J Polit Sci 59(2):495\u0026ndash;510. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/ajps.12116\u003c/span\u003e\u003cspan address=\"10.1111/ajps.12116\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbadie A, Gardeazabal J (2003) The economic costs of conflict: a case study of the Basque Country. Am Econ Rev 93(1):113\u0026ndash;132. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1257/000282803321455188\u003c/span\u003e\u003cspan address=\"10.1257/000282803321455188\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRelat\u0026oacute;rio anual de seguran\u0026ccedil;a de barragens de minera\u0026ccedil;\u0026atilde;o 2023. ANM, ANM - Ag\u0026ecirc;ncia Nacional de Minera\u0026ccedil;\u0026atilde;o, Bras\u0026iacute;lia (2023) Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gov.br/anm/pt-br/assuntos/barragens/relatorios-anuais-de-seguranca-da-barragens-de-mineracao-2/relatorio-anual-2023-final-1.pdf\u003c/span\u003e\u003cspan address=\"https://www.gov.br/anm/pt-br/assuntos/barragens/relatorios-anuais-de-seguranca-da-barragens-de-mineracao-2/relatorio-anual-2023-final-1.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAouadi A, Marsat S (2018) Do ESG controversies matter for firm value? Evidence from international data. J Bus Ethics 151:1027\u0026ndash;1047. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10551-016-3213-8\u003c/span\u003e\u003cspan address=\"10.1007/s10551-016-3213-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBatista ICAA, Firme VAC (2025) Economic impacts of Brazil\u0026rsquo;s Mariana dam disaster on different levels of neighbourhood. Spat Economic Anal 20(4):728\u0026ndash;750. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/17421772.2025.2502355\u003c/span\u003e\u003cspan address=\"10.1080/17421772.2025.2502355\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBiazoli L, de \u0026Aacute;vila ES, de Oliveira IRC (2024) Combining cluster analysis with synthetic control for evaluating economic impacts of the dam breach in Mariana, Brazil. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00181-024-02627-7\u003c/span\u003e\u003cspan address=\"10.1007/s00181-024-02627-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Empirical Economics\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarrillo B, da Mata D, Emanuel L, Lopes D, Sampaio B (2020) Avoidable environmental disasters and infant health: evidence from a mining dam collapse in Brazil. Health Econ 29(12):1\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/hec.4151\u003c/span\u003e\u003cspan address=\"10.1002/hec.4151\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCastro LSD, Almeida ES (2019) Desastres e desempenho econ\u0026ocirc;mico: avalia\u0026ccedil;\u0026atilde;o do impacto do rompimento da Barragem de Mariana. Geosul 34(70):406\u0026ndash;429. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5007/2177-5230.2019v34n70p406\u003c/span\u003e\u003cspan address=\"10.5007/2177-5230.2019v34n70p406\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCastro LSD, Almeida E (2023) Avalia\u0026ccedil;\u0026atilde;o do desastre de Brumadinho no desempenho econ\u0026ocirc;mico de Minas Gerais. Nova Economia 33(2):421\u0026ndash;447. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1590/0103-6351/7658\u003c/span\u003e\u003cspan address=\"10.1590/0103-6351/7658\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChang SE, Brown C, Handmer J, Helgeson J, Kajitani Y, Keating A, Roa-Henriquez A (2022) Business recovery from disasters: lessons from natural hazards and the COVID-19 pandemic. Int J Disaster Risk Reduct 80:103191. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ijdrr.2022.103191\u003c/span\u003e\u003cspan address=\"10.1016/j.ijdrr.2022.103191\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen G (2020) Assessing the financial impact of natural disasters on local governments. Public Budg Finance 40(1):22\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/pbaf.12245\u003c/span\u003e\u003cspan address=\"10.1111/pbaf.12245\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDalcero K, Rover S, Ferreira DDM (2024) Um mar de lama: h\u0026aacute; impacto no valor de mercado da Vale S/A p\u0026oacute;s desastres? Revista de Administra\u0026ccedil;\u0026atilde;o. Contabilidade e Economia da Fundace 15(2). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.13059/racef.v15i2.1181\u003c/span\u003e\u003cspan address=\"10.13059/racef.v15i2.1181\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDuarte AP (2008) Classifica\u0026ccedil;\u0026atilde;o das barragens de conten\u0026ccedil;\u0026atilde;o de rejeitos da minera\u0026ccedil;\u0026atilde;o e de res\u0026iacute;duos industriais no estado de Minas Gerais em rela\u0026ccedil;\u0026atilde;o ao potencial de risco. MSc dissertation, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFabr\u0026iacute;cio SA, Ferreira DDM, Borba JA (2021) A panorama of Mariana and Brumadinho disasters: what do we know so far? REAd. Revista Eletr\u0026ocirc;nica de Administra\u0026ccedil;\u0026atilde;o 27:128\u0026ndash;152. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1590/1413-2311.310.102806\u003c/span\u003e\u003cspan address=\"10.1590/1413-2311.310.102806\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFama EF (1970) Efficient capital markets: a review of theory and empirical work. J Finance 25(2):383\u0026ndash;417\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFama EF (1991) Efficient capital markets: II. J Finance 46(5):1575\u0026ndash;1617\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFernandes GW, Goulart FF, Ranieri BD et al (2016) Deep into the mud: ecological and socio-economic impacts of the dam breach in Mariana, Brazil. Natureza \u0026amp; Conserva\u0026ccedil;\u0026atilde;o 14(2):35\u0026ndash;45. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ncon.2016.10.003\u003c/span\u003e\u003cspan address=\"10.1016/j.ncon.2016.10.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFoga\u0026ccedil;a PACS, Raeder FT, Marques JAVC (2023) An\u0026aacute;lise dos impactos dos acidentes ambientais de Mariana e Brumadinho nas a\u0026ccedil;\u0026otilde;es da Mineradora Vale. REUNIR: Revista de Administra\u0026ccedil;\u0026atilde;o, Contabilidade e Sustentabilidade 13(2):1\u0026ndash;18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18696/reunir.v13i2.1426\u003c/span\u003e\u003cspan address=\"10.18696/reunir.v13i2.1426\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHalkos G, Zisiadou A (2021) Can we hedge an investment against a potential unexpected environmental disaster? Econ Disasters Clim Change 5:355\u0026ndash;365. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s41885-021-00085-4\u003c/span\u003e\u003cspan address=\"10.1007/s41885-021-00085-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang Q, Li Y, Lin M, McBrayer GA (2022) Natural disasters, risk salience, and corporate ESG disclosure. J Corp Finance 72:102152. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jcorpfin.2021.102152\u003c/span\u003e\u003cspan address=\"10.1016/j.jcorpfin.2021.102152\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIBRAM - Instituto Brasileiro de Minera\u0026ccedil;\u0026atilde;o (2024) Contribui\u0026ccedil;\u0026atilde;o da ind\u0026uacute;stria da minera\u0026ccedil;\u0026atilde;o para o G20: construindo um mundo justo e um planeta sustent\u0026aacute;vel. IBRAM, Bras\u0026iacute;lia. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://ibram.org.br/wp-content/uploads/2024/11/IBRAM_Contribuicao-da-Industria-da-Industria-da-Mineracao-para-o-G20-Brasil-2024-2-1.pdf\u003c/span\u003e\u003cspan address=\"https://ibram.org.br/wp-content/uploads/2024/11/IBRAM_Contribuicao-da-Industria-da-Industria-da-Mineracao-para-o-G20-Brasil-2024-2-1.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKeerthiratne S, Tol RSJ (2017) Impact of natural disasters on financial development. Econ Disasters Clim Change 1:33\u0026ndash;54. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s41885-017-0002-5\u003c/span\u003e\u003cspan address=\"10.1007/s41885-017-0002-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKrupa J, Jones C (2013) Black swan theory: applications to energy market histories and technologies. Energy Strategy Reviews 1(4):286\u0026ndash;290. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.esr.2013.02.004\u003c/span\u003e\u003cspan address=\"10.1016/j.esr.2013.02.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKuruc K (2022) Are IMF rescue packages effective? A synthetic control analysis of macroeconomic crises. J Monet Econ 127:38\u0026ndash;53. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jmoneco.2022.02.002\u003c/span\u003e\u003cspan address=\"10.1016/j.jmoneco.2022.02.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLSEG (2025) LSEG ESG Scores: methodology. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.refinitiv.com/content/dam/marketing/en_us/documents/methodology/esg-scores-methodology.pdf\u003c/span\u003e\u003cspan address=\"https://www.refinitiv.com/content/dam/marketing/en_us/documents/methodology/esg-scores-methodology.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (accessed 12 Dec 2025)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaran R (2023) Do sovereign catastrophe bonds improve fiscal resilience? An application of synthetic control method to Mexico (preprint). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.21203/rs.3.rs-2965725/v1\u003c/span\u003e\u003cspan address=\"10.21203/rs.3.rs-2965725/v1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Research Square\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMansur R (2025) Mariana, 10 anos: ningu\u0026eacute;m foi condenado por trag\u0026eacute;dia que matou 19 pessoas, destruiu comunidades e contaminou Rio Doce. G1 (3 Nov 2025). Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://g1.globo.com/mg/minas-gerais/noticia/2025/11/03/mariana-10-anos-ninguem-foi-condenado-por-tragedia-que-matou-19-pessoas-destruiu-comunidades-e-contaminou-rio-doce.ghtml\u003c/span\u003e\u003cspan address=\"https://g1.globo.com/mg/minas-gerais/noticia/2025/11/03/mariana-10-anos-ninguem-foi-condenado-por-tragedia-que-matou-19-pessoas-destruiu-comunidades-e-contaminou-rio-doce.ghtml\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMart\u0026iacute;nez-Ferrero J, Garc\u0026iacute;a-S\u0026aacute;nchez IM, Cuadrado-Ballesteros B (2015) Effect of financial reporting quality on sustainability information disclosure. Corp Soc Responsib Environ Manag 22(1):45\u0026ndash;64. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/csr.1330\u003c/span\u003e\u003cspan address=\"10.1002/csr.1330\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMasys AJ (2012) Black swans to grey swans: revealing the uncertainty. Disaster Prev Manage 21(3):320\u0026ndash;335. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/09653561211234507\u003c/span\u003e\u003cspan address=\"10.1108/09653561211234507\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMenezes C, Geiss C, Tressler J (1980) Increasing downside risk. Am Econ Rev 70:921\u0026ndash;932\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRhee SG, Wu FH (2020) Conditional extreme risk, black swan hedging, and asset prices. J Empir Finance 58:412\u0026ndash;435. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jempfin.2020.07.002\u003c/span\u003e\u003cspan address=\"10.1016/j.jempfin.2020.07.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRietz T (1988) The equity risk premium: a solution. J Monet Econ 22(1):117\u0026ndash;131\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRocha CH, Vasconcelos JG (2023) Barragens de Mariana e Brumadinho e pre\u0026ccedil;o da a\u0026ccedil;\u0026atilde;o da Vale: um estudo de evento. Revista Estudos e Pesquisas em Administra\u0026ccedil;\u0026atilde;o 7(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.30781/repad.v7i1.14785\u003c/span\u003e\u003cspan address=\"10.30781/repad.v7i1.14785\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRosenbaum PR (2007) Interference between units in randomized experiments. J Am Stat Assoc 102(477):191\u0026ndash;200. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1198/016214506000001112\u003c/span\u003e\u003cspan address=\"10.1198/016214506000001112\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoy AD (1952) Safety first and the holding of assets. Econometrica 20:431\u0026ndash;449\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSampaio B, Azevedo R, Azuaga FL (2016) Impacto de mudan\u0026ccedil;a regulat\u0026oacute;ria sobre os pre\u0026ccedil;os dos ativos das empresas do setor el\u0026eacute;trico brasileiro. Anais do 44\u0026ordm; Encontro Nacional de Economia, Foz do Igua\u0026ccedil;u, PR, Brazil\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSimonato TC (2017) Proje\u0026ccedil;\u0026atilde;o dos impactos econ\u0026ocirc;micos regionais do desastre de Mariana-MG. MSc dissertation, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSouza PA (2019) Open data could have helped us learn from another mining dam disaster. Sci Data 6:54\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTaleb NN (2007) The Black Swan: the impact of the highly improbable. Penguin Books, London\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWMTF - World Mine Tailings Failures (2019) World mine tailings failures (database). Available at: https://worldminetailingsfailures.org (accessed 1 Nov 2024)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWoods DD (2006) Essential characteristics of resilience. In: Hollnagel E, Woods DD, Leveson N (eds) Resilience engineering: concepts and precepts. Ashgate, Aldershot, pp 21\u0026ndash;34\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Footnotes","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e Approximately US\u003cspan\u003e$\u003c/span\u003e 27.5\u0026nbsp;billion, converted at the year-end 2024 PTAX rate of R\u003cspan\u003e$\u003c/span\u003e 6.1923 per US\u003cspan\u003e$\u003c/span\u003e 1.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e Original values (R\u003cspan\u003e$\u003c/span\u003e of 2022): direct losses R\u003cspan\u003e$\u003c/span\u003e 151.9 bn, R\u003cspan\u003e$\u003c/span\u003e 297.9 bn, R\u003cspan\u003e$\u003c/span\u003e 498.2 bn; total net effects R\u003cspan\u003e$\u003c/span\u003e 81.7 bn, R\u003cspan\u003e$\u003c/span\u003e 146.3 bn, R\u003cspan\u003e$\u003c/span\u003e 256.0 bn (Batista \u0026amp; Firme, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2025\u003c/span\u003e)\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"economics-of-disasters-and-climate-change","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"edac","sideBox":"Learn more about [Economics of Disasters and Climate Change](http://link.springer.com/journal/41885)","snPcode":"41885","submissionUrl":"https://submission.nature.com/new-submission/41885/3","title":"Economics of Disasters and Climate Change","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Mariana disaster, tailings dam collapse, synthetic control, stock returns, disaster finance, market efficiency","lastPublishedDoi":"10.21203/rs.3.rs-8894226/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8894226/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study examines how the Fundão tailings dam collapse (Mariana, Brazil) was priced in Vale S.A.’s equity returns, and whether the observed dynamics are consistent with rapid adjustment under the Efficient Market Hypothesis (EMH) when the shock has ‘black swan’ features. We apply the synthetic control method to construct a transparent counterfactual path for Vale’s semiannual cumulative stock returns over 2011.2-2017.2. To address potential spillovers and comparator contamination, we estimate three donor-pool specifications: DP1 (Brazilian and international peers), DP2 (excluding Brazilian firms), and DP3 (excluding Brazilian firms and BHP Group). DP1 delivers the strongest pre-disaster fit and is adopted as the baseline. The results show a negative return response in the disaster semester (2015.2), consistent with an immediate shock, but the most informative pattern arises afterward: Vale underperforms its synthetic counterfactual persistently in 2016.1-2017.2. This sustained post-event gap suggests delayed market adjustment as legal, regulatory, and reputational uncertainties unfold beyond the initial news window. Placebo-style comparisons support the interpretation that Vale’s post-disaster divergence is not a mere artifact of random fluctuations under strong pre-event fit. Overall, the study contributes to disaster finance by providing firm-level evidence on the persistence of disaster-related repricing and by highlighting governance, transparency, and disaster-risk management as relevant mechanisms for limiting valuation losses under extreme events.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eJEL classification: \u003c/strong\u003eG14; Q54; Q56; L72; C23\u003c/p\u003e","manuscriptTitle":"How Do Equity Markets Price Disaster Risk? Evidence from Vale after the Mariana Disaster","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-06 08:30:19","doi":"10.21203/rs.3.rs-8894226/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-04-23T11:11:10+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-21T13:52:24+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-06T12:49:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"273511182971124227592940419653531069950","date":"2026-03-17T06:31:17+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"198209263841625268390022360733992624626","date":"2026-03-03T11:09:35+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-03T07:20:36+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-23T18:46:12+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-20T03:08:52+00:00","index":"","fulltext":""},{"type":"submitted","content":"Economics of Disasters and Climate Change","date":"2026-02-16T14:55:21+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"economics-of-disasters-and-climate-change","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"edac","sideBox":"Learn more about [Economics of Disasters and Climate Change](http://link.springer.com/journal/41885)","snPcode":"41885","submissionUrl":"https://submission.nature.com/new-submission/41885/3","title":"Economics of Disasters and Climate Change","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"f353d97e-698e-4373-82ae-8055a8d7fb83","owner":[],"postedDate":"March 6th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[],"tags":[],"updatedAt":"2026-04-23T11:25:39+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-06 08:30:19","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8894226","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8894226","identity":"rs-8894226","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}