Temperature shocks and stock returns in Brazil: a spatio-temporal approach to estimating the risk premium

This study investigates how long-term climate dynamics affect asset pricing in an emerging market context by introducing a novel temperature risk factor derived from a spatio-temporal decomposition of temperature data in Brazil. Unlike conventional models, our approach distinguishes between predictable climate trends and unanticipated shocks by using geostatistical techniques to isolate the innovations of the temperature trend. Empirical results show that temperature shocks carry a statistically significant and positive risk premium in the Brazilian stock market. Most industry portfolios exhibit negative exposure, indicating that these sectors serve as a hedge for unexpected increases in temperature trends. These findings are particularly relevant for asset managers, policymakers, and ESG-focused investors, as they highlight sector-specific vulnerabilities and the systemic nature of climate risks in developing economies. By incorporating spatial heterogeneity and climate dynamics into standard pricing frameworks, this study advances the field of climate finance and provides a replicable methodology for enhancing climate-informed investment strategies.