Doctor of Philosophy
School of Business
This thesis on climate change disclosures comprises three empirical studies in which computerised textual analysis is used to develop a climate change disclosure score linked to asset pricing, analyst forecasts, and carbon emissions from the different perspectives of investors, analysts, and corporate management in capital markets.
The first study examines the role of climate change disclosures in predicting stock returns. I find that a long-short trading portfolio sorted on climate change disclosures earns an annualised risk-adjusted return of 9.11% and 6.96% in Australia and the US respectively, after controlling for common risk factors. The Fama-MacBeth regressions consistently show that climate change disclosures negatively predict stock returns in both countries. I also find that lower climate change disclosure scores are associated with greater return volatility in firms, indicating higher future risk. This suggests that investors face greater uncertainty with climate risk assessment for firms with low disclosures, and hence demand higher returns as compensation.
The second study examines whether analysts incorporate information contained in climate change disclosures into their earnings forecasts. The results show that increased disclosure scores are associated with reduced forecast error and dispersion, where the relationship is more pronounced over shorter-term forecasting periods among firms that support the Task Force on Climate-related Financial Disclosures (TCFD). Additionally, TCFD category-level climate change disclosures under the categories of Governance, Strategy, and Risk Management specifically help to improve forecasts. The main results remain robust after additional tests including sample selection endogeneity, country-level analysis, and alternative forecast measures.
The third study explores corporate environmental accountability by examining how carbon emissions affect voluntary climate change disclosures based on TCFD principles. The results show that firms with higher levels of carbon emissions disclose more climate change information. This relationship is stronger in firms belonging to carbon intensive industries, such as energy, materials, and utilities. I also investigate this relationship using a category-level disclosure measure, finding that carbon emissions drive disclosures under Strategy, Risk Management, and Metrics and Targets categories. Overall, the evidence shows that high carbon emitting firms appear to discharge their corporate accountability by increasing climate change disclosure, consistent with legitimising their potentially unethical actions and submitting to stakeholder and societal pressure.
Overall, climate change disclosures convey useful information to market participants, including investors and analysts for decision-making, and reflect corporate managers’ accountability to improve information transparency. Specifically, climate change disclosure scores predict negative stock returns on average, because investors demand higher return as a compensation for firms with lower levels of disclosure, indicating greater uncertainty in assessing firms’ climate risk exposure. Additionally, disclosure scores are negatively associated with analyst forecast errors and dispersions. This suggest that financial analysts can infer useful information contained in climate change disclosures to improve earnings forecasts, specifically through information quality assessment and an informed understanding of the financial implications of climate risks from category-level disclosures. Lastly, higher carbon emissions increase climate change disclosures, implying that corporate managers are likely to increase disclosures to improve a firm's environmental image and alleviate social pressure. These findings collectively have important implications for understanding the role of climate change disclosures in capital markets.
Ding, Dong, Essays on Firm-level Climate Change Disclosure, Doctor of Philosophy thesis, School of Business, University of Wollongong, 2023. https://ro.uow.edu.au/theses1/1643
FoR codes (2008)
080107 Natural Language Processing, 150106 Sustainability Accounting and Reporting, 150201 Finance
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.