Assessing the Impact of Sustainability on Fund Flows: An Excess Information Approach and US-Based Case Study

Abstract

Motivated by studies of the interplay between environmental, social, and governance (ESG) factors and fund flows, the authors demonstrate an alternative approach for analyzing the impact of sustainability on fund flows and present a case study applying their approach to US-domiciled equity mutual funds. In particular, they use supervised machine learning techniques to isolate excess information available within ESG drivers. This approach is chosen instead of commonly used regression approaches to control for multicollinearity and address the known correlation of sustainability to other factors. With respect to the case study undertaken, the authors find that when controlling for nonlinear drivers and interactions between input factors (past fund flows, past returns and variances, fees and other fund-level variables), sustainability provides little excess information for predicting fund flows. This result is seen through a comparison of two predictive models with a robust benchmark that uses only features unrelated to ESG/sustainability and has only marginally worse predictive power compared to a model built using all available data.