Sustainability Disclosure on Scope 3 GHG Emissions: Evidence from Oil & Gas Producers

Can transparency trigger a backlash for heavy emitters?

Maretno Harjoto, Andreas Hoepner, and Fabiola Schneider analyse 48 oil and gas firms worldwide during the COVID-19 demand shock in « Sustainability Disclosure on Scope 3 GHG Emissions: Evidence from Oil & Gas Producers ».

They find that firms revealing the « elephant in the room » by announcing Scope 3 emissions led investors to mark them down and faced negative stock returns relative to non-disclosers. Their finding include:

• Shareholders reacted as if the disclosed Scope 3 exposure made the firms more risky: sustainability transparency can backfire if it highlights major risks without a clear mitigation plan.
• The underperformance was concentrated among firms reporting « Use of Sold Products » emissions (the CO2 from customers burning the fuel), as it highlights a core risk for oil and gas businesses.
• The effect persisted after controlling for self-selection: no similar stock drop occurred for firms disclosing only Scope 1 or 2 emissions, nor in a comparable high-emission industry (airlines) during the same shock.
• The study estimates a roughly -0.2% daily return difference for disclosing firms, with the impact highly significant statistically.
• European oil and gas companies were far more likely to voluntarily disclose Scope 3 data, suggesting these firms responded to institutional pressure to maintain legitimacy despite potential market backlash.

This research shows climate transition risk is being priced in: when a company exposes the full scale of its carbon liabilities, it can materially affect valuation under certain conditions.

Investors perceive mere disclosure as « too little, too late » in the absence of concrete transition strategies. While transparency is important, it needs to be paired with credible decarbonisation plans to avoid sending a negative signal.

In spite of some robustness tests, the analysis is limited to one industry and to voluntary disclosures, which came with data gaps. The findings may not generalise well beyond this setting.