Pricing climate risk in U.S. equities

: evidence from tail risk and stress regimes

Abstract

This study shows how climate risks are priced in U.S. equity markets and whether the pricing varies across different economic-distress regimes. I use nine machine-learning-derived textual factors to capture physical and transition risks, and employ conditional bivariate portfolio sorts to estimate risk premia. A Hidden Markov Model (HMM) categorizes market distress into low, medium, and high regimes for testing regime-based pricing. Portfolios with high exposure to Physical Risk Index (PRI) generate persistent negative alphas during stress periods. Conversely, those with high exposure to transition risks such as U.S. Climate Policy (USCP) and Interna- tional Summits (IS) yield positive premia in all regimes. A new composite textual factor yields larger returns, highlighting a benefit from combined narrative exposures. Narrative indicators of climate stress confirm distinct pricing channels, while tail-risk analysis shows that exposures to the PRI drive losses in severe scenarios, whereas exposures to the Transition Risk Index (TRI) are resilient. First, these insight inform and urge for robust dynamic risk-management strategies, but also suggest how climate-related narratives are shaping investors’ perceptions and pricing dynamics.

Date of Award	21 Oct 2025
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Bruno Gerard (Supervisor)