ESG Risk Assessment of Mortgage Loans

Work thumb

Views: 23

  • Title: ESG Risk Assessment of Mortgage Loans: Mass Estimation of Energy Performance Certificates in Situations with Limited Information and Financial Resources
  • Author(s): Timotej Jagric, Aljaž Herman
  • Publisher: Common Ground Research Networks
  • Collection: Common Ground Research Networks
  • Series: On Sustainability
  • Journal Title: The International Journal of Sustainability Policy and Practice
  • Keywords: Energy Performance Certificates, Banking and ESG Risks, Machine Learning Classification, Bagged Trees
  • Volume: 21
  • Issue: 1
  • Date: December 09, 2024
  • ISSN: 2325-1166 (Print)
  • ISSN: 2325-1182 (Online)
  • DOI: https://doi.org/10.18848/2325-1166/CGP/v21i01/151-171
  • Citation: Jagric, Timotej, and Aljaž Herman. 2024. "ESG Risk Assessment of Mortgage Loans: Mass Estimation of Energy Performance Certificates in Situations with Limited Information and Financial Resources." The International Journal of Sustainability Policy and Practice 21 (1): 151-171. doi:10.18848/2325-1166/CGP/v21i01/151-171.
  • Extent: 21 pages

All Rights Reserved

Copyright © 2024, Common Ground Research Networks, All Rights Reserved

Abstract

Green behavior and the desire to act sustainably are two perspectives that are extremely important from an environmental point of view. In the realm of mortgage products, there has been a significant emphasis on the environmental aspect of environmental, social, and governance considerations. One area where we can act in this way is the property market. Here, energy performance certificates (EPCs), which determine the energy class of a building, play an important role, especially from the point of view of banks. Banks see the contribution of EPCs as an added value in assessing the riskiness of a building, which has a major impact when we talk about lending, investments, and other banking aspects. In our research, we tested twenty-eight different models, looking for the one that would classify buildings into energy classes as accurately as possible. The bagged trees model proved to have the best classification power, achieving 72.4% accuracy.