ESG in Asset Management: Is SFDR Effective in Segmenting Funds with Different Sustainability Profiles?

Abstract

This thesis investigates the effectiveness of the European Union’s Sustainable Finance Disclosure Regulation (SFDR) in differentiating European mutual funds and ETFs based on their sustainability profiles. The SFDR mandates asset managers to provide clear indications concerning the sustainability objectives of their funds. In particular, fund providers operating in Europe must classify their financial instruments as either Article 6, Article 8, or Article 9, depending on the specific social and environmental objectives specified in their pre-contractual disclosures. However, contrasting evidence has been found regarding whether this regulatory-mandated segmentation is reflected in the actual ESG performance of the funds’ portfolios. To advance this area of research, we carry out statistical analyses using ESG data extracted from the LSEG Workspace platform. The thesis is structured into three sections.

The first section will provide an overview of the concept of ESG in financial markets. By analysing its emergence and relevance to the current asset management industry, we set the stage for assessing the SFDR and the evidence regarding its effectiveness. This will set the stage for the analytical section of this thesis, which is further divided into two distinct methodological approaches.

In the first part, we analyse descriptive statistics to identify significant differences in ESG performance across funds with different SFDR labels. We apply Welch’s ANOVA and Games-Howell Tests to evaluate whether there are statistically significant differences in funds’ sustainability performance between the groups, testing our first hypothesis. The second part applies cluster analysis to ESG performance data. Using the K-Means clustering algorithm and confusion matrices, we assess whether funds naturally cluster into three groups corresponding to the observed SFDR segmentations, testing our second hypothesis.

Our findings produce unexpected results. We find that distinct SFDR affiliations do not cause investment funds’ sustainability profiles to differ at statistically significant levels. In addition, we find that our results change drastically based on the specific ESG metrics analysed. Finally, we observe that the clustering results based on sustainability parameters are not comparable to the SFDR-based partitions observed in the real world.

These results indicate that the SFDR is ineffective in inducing asset managers to accurately self-select investment funds based on their sustainability profiles.