The following perspective is from Philippe Jordan, President of Capital Fund Management (CFM). For his full insights on ESG investing, go here.
It is clear that ESG data presents as both the biggest impediment to security selection, and (probably) the most commercial opportunity. ESG also fits nicely into the contemporary thinking that traditional indicators could (and should) be augmented with non-traditional data in forming a more holistic view of firm risk and value.
The story of ESG is in large part the story of alternative data, with the unstructured nature of data an opportunity for systematic quant investors being uniquely positioned to uncover and discern between what information could be alpha generating or not. As it stands, managers could either tap into the vast choice of commercially available data sources (while being cognizant of the hazards of this approach), or employ their own skill to scrutinize alternative data. This will, either way, imply a cost to those managers willing to make an investment in this research.
It is, however, clear that many Hedge Funds remain uneager to embark on such an endeavor. A survey by Unigestion reveals that more than half of European-based hedge funds are reluctant about ESG integration, compared to nearly 70% of US-based hedge funds.
A main concern for the lack of interest, is the unease about the quality of data. That data should be improved is a widely agreed objective, advocated by the finance industry and non-governmental organizations. A few organizations are making strides into setting the pace for data quality and comparability. The Task Force on Climate-related Financial Disclosures (TCFD), which helps companies identify climate-related risks, is one. There are, furthermore, various policy initiatives driving an increase in corporate disclosures related to ESG issues. According to the Governance & Accountability Institute, 85% of S&P 500 Companies published some form of sustainability report in 2017. This is up from the 20% that published these types of report in 2011, despite different levels of transparency.
There are claims that machine learning, big data and artificial intelligence are enabling better analysis of ESG data. Despite the benefits and opportunities this presents in a systematic framework, this claim should be met with some caution. Machine learning is not always uniquely positioned to tackle all problems in finance. A recent paper by Arnott et al. highlights one of the “crucial limitations” of machine learning, namely “data availability”, since machine learning applications work best when there exists a vast amount of data. While the amount of data that could pertain to ESG-like issues has increased, one is susceptible to overfitting and finding spurious relationships with data that are not relevant, or material.
Another barrier often cited for ESG investment practices is the limited understanding of the material ESG issues affecting firms. It is likely that firms will allocate more resources and expertise if they are truly committed to ESG investment.
Learn more about the what, why, and decisively, the how of ESG investing from a quant perspective.
