Journal of Banking & Finance 2026
Understanding Corporate Emissions Data: Implications for Research Design
Abstract
Carbon emissions data are widely used in environmental and related research. This paper examines several empirical properties of greenhouse gas (GHG) emissions data that are relevant to research design and interpretation. First, the rapid increase in publicly listed firms with available GHG emissions data has changed sample composition over time. Second, emissions data are largely vendor-estimated rather than disclosed by firms. Compared with disclosure-linked observations, estimated emissions exhibit greater cross-vendor disagreement, strong associations with USD-denominated firm sales and exchange rate fluctuations against USD, and substantially less within-industry variation in emissions intensity. Third, only 5.3% of Compustat firm-year observations disclose their own GHG emissions data. These findings suggest that empirical results can be sensitive to emissions data types, measurement choices, control variable selection, and sample composition. We provide research-design recommendations to account for these properties and demonstrate their value through empirical tests.
- DOI
- 10.1016/j.jbankfin.2026.107822
- Pages
- 107822
- Language
- en
- Sources
- crossref openalex