Title: The environmental impact of green bonds: Data-driven insights through statistical learning Authors:  Christina Erlwein-Sayer - University of Applied Sciences HTW Berlin (Germany) [presenting]
Abstract: Green bonds have become a transformative financial instrument aimed at financing initiatives with positive environmental outcomes. We investigate the efficacy of corporate green bond issuances in improving environmental performance at an individual issuer level, providing insights into issuer characteristics and factors influencing environmental impact. Based on a controlled interrupted time-series (CITS) model, green bonds and conventional bonds are compared to estimate the individual effects of green bonds on issuers' environmental (E) scores. This is followed by a two-stage statistical analysis. In the first stage, a random forest model is built to classify the key factors influencing environmental performance improvements, revealing that bond attributes, notably Climate Bonds Initiative certification, had minimal predictive power. In the second stage, we employ a generalized additive model to capture non-linear relationships between green bond effects and additional explanatory variables. Our results show that green bonds yield variable impacts on E scores across issuers. Non-linear relationships are observed, especially for bond value and issuer characteristics, affecting E-score changes. Despite some green bonds not affecting performance significantly, the study underlines the positive environmental impact of green bonds overall, urging policymakers to support green financing, yet highlighting issuer characteristics as more influential than the bonds themselves.