Creating a Systematic ESG Scoring System: Purpose

Table of Links

• Abstract and Introduction
• Related Works
• Purpose
• Methods
• Results
• Discussion
• Conclusion and Bibliography

3. Purpose

The purpose of this project was to create a systematic ESG rating system that gives executives and outsiders a more balanced and representative view of a company’s practices for greater social responsibility. To do this, a machine-learning algorithm was created using social network data to quantitatively evaluate ESG. Social network data was used instead of self-reported filings since it can provide various outsider perspectives on issues people feel a corporation should address. By directly showcasing public opinion, it can remove the bias of self-reporting and help executives create more targeted initiatives for meaningful change. Furthermore, a data-driven system can provide ESG ratings for companies without coverage.

To test the predictive power of the proposed system, the correlation as well as the mean absolute average error (MAAE) were measured against current ESG ratings. This can help determine whether the system is viable for rating prediction. However, potential constraints include limited access to high volumes of social network data, the accuracy of NLP algorithms, and limited computational resources.

The contributions of this work can be summarized as follows:

 It gives a real-time social-sentiment ESG score that highlights how people feel regarding a company’s practices. This can give executives a way to monitor the ESG health of their organization. It also shows which areas the people feel need the most change, and this can help target executive initiatives to be more effective.

 It provides a full-stack method for gathering real-time ESG data and converting it into a comprehensive score. This allows for the readily available creation of initial ESG ratings that can be used either directly by investors to ensure they are making socially conscious investments (especially for non-rated companies) or by ESG rating agencies to scale up coverage.

 The proposed approach utilizes multiple social networks for score prediction. Most papers about ESG social network analysis typically hyperfocus on one specific network such as Twitter or the News (Sokolov et al., 2021). This paper seeks to combine them while also adding other under-analyzed social networks (i.e., LinkedIn, Wikipedia).