Abstract How do equity markets price low-probability, high-severity environmental disasters? We study this question through Vale S.A., the firm at the center of two of the largest tailings-dam failures on record within less than four years: Mariana (November 2015), at a dam jointly owned with BHP, and Brumadinho (January 2019), under Vale’s sole control. Using monthly data from 2010 to 2025, we build a transparent counterfactual for Vale’s stock returns with the synthetic control method and its augmented version, isolating each event in its own window and assessing significance through in-space and in-time placebo tests and a complementary factor model. The two disasters were priced very differently. For Mariana, once Vale is compared with a credible counterfactual, no abnormal return survives: the firm’s later outperformance reflects its heavier exposure to the iron-ore cycle, not the disaster. For Brumadinho, the market repriced Vale sharply in the event month and then corrected within a year, a pattern closer to semi-strong market efficiency than to delayed adjustment. Read together, the two cases support a conditional view of disaster pricing: a shock is priced not automatically, but in proportion to how clearly responsibility can be attributed to the firm and how severe the event is. The study brings firm-level, market-based evidence to disaster finance and points to governance, transparency, and clear attribution as conditions for accurate pricing.
Mendes et al. (Mon,) studied this question.