University of Twente Student Theses

Login

Climate Change Risk for Financial Institutions : Predicting Corporate Greenhouse Gas Emissions

Rothman, Ties (2023) Climate Change Risk for Financial Institutions : Predicting Corporate Greenhouse Gas Emissions.

[img] PDF
3MB
Abstract:In this research, we investigate the use of statistical analysis and machine learning methods to predict corporate greenhouse gas emissions. We trained and tested these models on corporations that disclose emission data, aiming to create models applicable to corporations that do not disclose this information. Due to a scarcity of Environmental, Social, and Governance (ESG)-related data, we used financial, geographical, and sector classification data as predictor variables. We applied a log transformation to both the predictor and output variables. Multiple imputation was employed to handle missing data, thereby enlarging our dataset while preserving underlying variable distributions. We evaluated the models in three rounds of testing: first on the imputed data, then on baseline data, and finally on baseline data after correcting for log transformation bias. In the log-transformed feature space, the models accurately predict corporate greenhouse gas emissions. However, in the original feature space, they fail to provide accurate predictions. Our findings suggest that the models struggle with the complexity of the data and do not generalize well. For more accurate predictions, additional ESG-related data, as well as information on production processes, materials, and other physical assets, are needed.
Item Type:Essay (Master)
Clients:
Zanders, Utrecht, Netherlands
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:85 business administration, organizational science
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:https://purl.utwente.nl/essays/97170
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page