Measuring Corporate Weather Exposure using Computational Linguistics

Abstract

The FASB and IFRS Foundation's International Sustainability Standards Board have called for measuring individual firms' exposure to weather, a fundamentally amorphous concept, as a first step toward quantifying the impact of environmental factors on financial reporting. This study builds a large-scale measure of individual firm exposure to weather using linguistic analysis of annual reports. Preliminary analyses suggest that weather is a determinant of our measure: e.g., it increases significantly after the firm gets hit by a severe storm. Despite being constructed from largely backward-looking mandated reports, our measure is forward looking in that it can predict variation in returns around future extreme weather events. Exposure to our measure is also priced as a risk factor, further establishing its forward-looking nature systematically in the cross-section. Our measure appears to reasonably capture a firm's business exposure to weather, thus showcasing the power of accounting to measure the economic impact of ESG factors.