A Multi-Design Approach to Hedging: The Case of Climate Risk

Abstract

Empirical research in finance involves numerous design choices, which can lead to substantial variation in results across plausible specifications. Rather than viewing these non-standard errors solely as a concern, this paper proposes a framework that systematically exploits such variation for hedge portfolio construction. The approach explores a large design space of candidate portfolios generated from alternative empirical specifications. It proceeds in three stages: estimating asset exposures to the hedging target, translating them into portfolio weights, and selecting and combining top-performing designs through a validation procedure. Applied to climate-risk hedging, the framework evaluates over 600,000 candidate designs and delivers out-of-sample results that significantly outperform traditional factor-mimicking portfolios across multiple climate-risk measures and sample periods.

Keywords: Design choices, non-standard errors, hedging strategies, climate risk

JEL: C52, C55, C58, G11, G17, Q54