Commentaries

June 2020 - Leveraging on ESG data and Artificial Intelligence to participate to the climate transition - Systematic Fund Manager's Comments

14 July 2020

Emmanuel Hauptmann

 Launch of the RAM Stable Climate Global Equities Fund

At RAM, ESG data integration in our systematic equity selection process has been in place for several years already. We now take this process further. As the climate change emergency continues to grow, we are convinced of our role as an asset manager to participate to the climate transition and provide investors with a differentiated solution to low-carbon investing. Thanks to tremendous advances made on our Machine Learning infrastructure, we have built a robust and active strategy with strong ESG standards, without compromising performance.

Intersection of fundamental financial and ESG data

We strongly believe the capture of persistence inefficiencies in the equity market resides in the interdependence between sustainability, return and risk characteristics. Our systematic approach takes a multi-dimensional angle to deliver “sustainable alpha”, leveraging on a myriad of ESG and non-ESG data sources to identify the most attractive risk-reward opportunities in each industry.

RAM AI’s proprietary ESG

ESG data sources we have been progressively integrating into our investment process over the past years carry value-added information from structured and unstructured data. Perfectly aware of the fact that more than three-quarter of new dataset relates to unstructured data, we have developed a state-of-the-art technique to extract information from finance text (i.e. news transcripts, earnings reports, etc) and transform them into quantitative “model friendly” features.

Natural Language Processing (NLP) remedying to the low frequency of ESG data

More conventional climate and ESG related data are at a low frequency and data providers would typically take days or weeks to react, while automatically analysing news flow helps us to identify the latest ESG related issues on companies and assess their wider impact. It is commonly accepted that real-world events reflected within unstructured data, e.g. financial news, earning calls, transcripts, financial reports, social media, etc, have a certain relationship to markets. NLP enables us to integrate inputs from these unstructured and qualitative data sources into our quantitative models. These inputs, which are complementary to our existing quantitative/structured data from analysts’ revisions, enrich the information set that our quantitative models consume.

On top of that, from the investment universe pre-processing perspective, NLP brings an “immediate” flag of controversies and ultimately a rapid exclusion of companies with strong negative ESG news.


Deep Learning models to address the multi-dimensionality problem

The important developments on the Machine Learning infrastructure permits us to process high dimensional data to make informed predictions, simultaneously integrating information across traditional fundamental financial, ESG and alternative data. We use an ensemble approach, consisting of a dozen of Machine Learning models with optimized hyperparameters. We regard Machine Learning techniques as a generalisation of traditional data processing techniques, and our research efforts are equally focused on testing models and controlling them by making sure they generalise and provide tangible results.


Compensating the portfolio carbon emissions by carbon certificates

In the RAM Stable Climate Global Equities Fund, carbon emissions of the portfolio (much lower than those of MSCI World Index) are compensated with carbon certificates issued through the Clean Development Mechanism of the UN. The projects targeted are biomass projects, which have a clear measurable impact on the environment and an auditing process at multiple levels. The cost of these carbon certificates is fully endorsed by RAM AI and not the Fund. In that way, we offer our investors a “pure” approach participating to the climate transition phase.

Further ESG integration across RAM AI’s equity product range

RAM AI’s further ESG data integration is being progressively deployed through our equity product range. We view this process as a continuation of the research efforts entertained over the last 5 years and the commitment and long-term alignment towards the SDGs. The cross-fertilisation of ideas within the RAM research team will lead to more ESG data integration and NLP techniques across the RAM product range to improve the alpha prediction.
RAM Emerging Markets Equities – Improved ESG rating with higher alpha stability
We have brought a sustainable alpha optimization layer to our Emerging Markets equities strategy, which gives a preference in the selection to companies with both high alpha and attractive sustainability. The additional sources of data relate to Agency ESG ratings and News Flow. The process of integrating the sustainability dimension into our alpha makes it more stable, and has the following positive impact on the Fund’s profile:

  • More sustainable All-Cap selection
  • Lower turnover and trading costs
  • Higher expected net alpha
  • Reduced active risk versus market-cap based benchmarks


As a result, the MSCI ESG rating of the portfolio is improved from BB to A (vs BB for MSCI Emerging Markets Index).

 

Direct access per fund to our latest Fund Manager's Comments: 

European Equities

Emerging Markets Equities

North American Equities

Global Sustainable Income Equities

Long /Short European Equities Fund

Long/Short Emerging Markets Equities

Long/Short Global Equities