Relating Corporate Sustainability Engagement with Environmental Quality: Empirical Evidence from India Using WebAI and Earth Observation Data

Environmental pollution remains a critical challenge in developing countries. This paper thus investigates the relationship between corporate sustainability engagement and the environmental conditions of the districts in which firms operate across districts in India. Operationalising organisational websites, we apply deep learning methods to construct firm-level sustainability engagement scores and aggregate them to the district level. We combine these indicators with earth-observation measures of air pollution and vegetation and estimate district-level regressions using spatial lag models to account for spatial dependencies. We find that more polluted districts host more sustainability-engaged firms in absolute terms. Yet, this association is largely absorbed by industrial concentration, population density, and institutional quality once controls are included. The relative share of sustainability-engaged firms shows no significant relationship with air quality. The study highlights the potential of large-scale web text analysis combined with complementary geodata for examining environmental sustainability engagement. A key limitation is the low availability of firm-level data, highlighting the need for improved data infrastructure for environmental and sustainability research in emerging economies.

Schmidt, Sebastian, Devika Jain, Johannes Dahlke, Jan Kinne, Robert Dehghan, Shaily Gandhi, Rockli Kim and S.V. Subramanian (2026), Relating Corporate Sustainability Engagement with Environmental Quality: Empirical Evidence from India Using WebAI and Earth Observation Data, ZEW Discussion Paper No. 26-043, Mannheim.