425 Figure 1 – CAR of responsible firms across event windows The results show that only a subset of events generated statistically significant direct effects at conventional levels. The clearest cases are Samarco (BHP Group) and Mount Polley (Imperial Metals). BHP reached a worst-window CAR of -29.97% in the [−21,+21] window, significant at the 1% level, while Imperial Metals reached -46.35% in [−14,+14], also significant at the 1% level. Brumadinho (VALE) also shows a significant direct effect, with a CAR of -20.71% in [−7,+7], significant at the 5% level. Other cases exhibit economically meaningful negative CAR without statistical significance. For example, Newmont and Teck show worst-window CAR of -28.65% and - 23.07%, respectively, but with p-values above 0.10. This indicates that large negative magnitudes may still arise even when the abnormal-return signal is not estimated precisely enough to reject the null hypothesis. Peer-firm results suggest that spillover effects are generally weaker than direct effects, but still relevant. The clearest case is Samarco, where peers exhibit an average CAR of - 33.44%, with 100% of peers showing negative CAR. A similar, though smaller, pattern appears in Mount Polley, Minas-Rio, and Río Sonora, where average peer CAR is negative and all or most peers are affected. In contrast, Brumadinho shows a strong direct impact on VALE but a positive average peer CAR, suggesting a more concentrated firm-specific effect. Overall, the evidence indicates that environmental disasters in mining may generate both direct valuation losses and moderate spillover effects on related firms. The strongest direct reactions are associated with catastrophic and highly visible events, while peer effects appear more diffuse and are better captured through average CAR and the proportion of negative peer responses than through individual significance alone. 3.2 Magnitude, significance, and event-window sensitivity An important result is that economic magnitude and statistical significance do not always coincide. Some events show large negative CAR but weak statistical significance. This is likely related to the high volatility of mining stocks, which are affected by commodity prices, macroeconomic conditions, exchange rates, and broader market sentiment. In such contexts, event-related price changes may be economically relevant but difficult to identify precisely in statistical terms. Figure 1 also shows that the response of responsible firms varies across event windows. For BHP, Imperial Metals, and VALE, CAR remains negative across windows, which suggests a stable market reassessment of firm value. In contrast, firms such as Rio Tinto and Vedanta exhibit sign changes across windows, indicating more ambiguous or less persistent reactions. For Newmont and Teck, CAR becomes more negative in longer windows, suggesting a more gradual adjustment process.
RkJQdWJsaXNoZXIy MTM0Mzk2