Impact of Climate Change on Firm Performance and Asset Pricing: Machine Learning and Text-Based Analyses

Climate change has become a central strategic and regulatory issue for firms worldwide (Barnett, 2023). Firms increasingly disclose climate-related information in annual, sustainability, and integrated reports, reflecting their climate awareness, risk perceptions, and transition strategies (Acheampong et al., 2025). However, it remains unclear whether these disclosures represent genuine strategic adaptation or symbolic communication, such as greenwashing.

These disclosures may also influence investors’ perceptions of climate risk and regulatory exposure, thereby affecting asset prices. Yet the extent to which climate-related information is incorporated into financial markets remains uncertain.

Using text-based analysis and natural language processing (NLP), this project extracts climate-related keywords from corporate reports to examine whether the intensity, tone, and content of disclosures are associated with firms’ financial performance, environmental performance, and risk management. It further investigates whether disclosures of quantitative greenhouse gas (GHG) emissions and carbon performance influence asset pricing and generate abnormal risk-adjusted returns. By applying machine learning to large-scale climate and financial datasets, the study explores whether climate-related information constitutes a priced risk factor in capital markets.

Research questions*:
1. How does the intensity and tone of climate change disclosure in annual reports relate to firms’ financial performance?
2. Do climate-related textual disclosures reflect genuine environmental performance or symbolic reporting?
3. Do corporate disclosures of quantitative GHG emissions influence asset pricing?
4. Can machine learning models using climate disclosure data generate abnormal risk-adjusted returns?
* The proposed research questions are a guide. Candidates are welcome to refine them or propose new questions, provided they align the objectives of this project.

Research methodology
The project will employ text-based analysis and secondary data collection. Data sources include firm annual reports, FAME, Bloomberg.
Contributions

The project will have significant contributions to literature, methodology, policy, regulation and practice on climate disclosure for investors and stakeholders.

The project will be supervised by a rich experienced supervision team, who have extensive experience in supervising PhD candidates and wide knowledge in areas of text-based analysis, asset pricing and climate change.

Supervisors:

Dr Thao Nguyen

Albert Acheampong

Dr Haozhe Su

References:
Barnett, M. (2023). Climate Change and Uncertainty: An Asset Pricing Perspective. Management Science. doi:https://doi.org/10.1287/mnsc.2022.4635.  
Albert, A., Mousavi, M., Gozgor, G. and Patrick, Y. (2025). The Impact of Text‐Based Financial Constraints on Stock Price Crash Risk: Evidence from the UK Firms. International Journal of Finance & Economics. doi:https://doi.org/10.1002/ijfe.3113.

Entrants must have a Master’s degree in accounting, finance, economics, business, management, operations research, or any other related area.

We particularly welcome candidates with strong quantitative skills in Stata, R and Python.