AI will not replace human-supervised ESG analysis

Rachida Mourahib of ODDO BHF underlines the contribution of technologies such as artificial intelligence to process data but also their limits.

The assessment of environmental, social and governance (ESG) criteria requires the processing of a considerable amount of data, which is added to the analysis of key financial figures. How is it possible to process this data in the most efficient way? And what criteria should you pay particular attention to when analyzing a company from a sustainability perspective? Interview with Rachida Mourahib, Global Head of ESG Research at ODDO BHF Asset Management, who gave a presentation at the Geneva Forum for Sustainable Investment (GFSI) held last week in Geneva.

How is the stock selection process for sustainable investing organized at ODDO BHM Asset Management?

Our ESG assessment approach is divided into three steps. First of all, there is an initial common exclusion base which leads managers to exclude certain securities from their investment universe. We exclude the controversial arms industry, the tobacco industry, coal-related activities or any company that is in proven violation of one of the 10 principles of the United Nations Global Compact.

“There is always a risk that a company will accommodate its upstream data a bit to be perceived more favorably, or less unfavourably.”

The second step is based on ESG integration. The aim here is to assess the extent to which corporate environmental, social and governance practices can have a material impact on a company’s business model and drivers of value, such as revenue growth. , margins, capital requirements or risks. This materiality differs from one sector to another and may relate to respect for the environment, the use of green energies, good social practices, balanced compensation for managers or even business ethics. ESG analysis integrated with financial analysis thus makes it possible to better assess the ESG risks of a company in investment decisions and in this way to generate long-term returns that are sustainable.

The third pillar is shareholder engagement. At ODDO BHF Asset Management, we do not limit ourselves solely to the exercise of voting rights at general meetings of shareholders. We also foster dialogue and engagement with companies, focusing on themes from sustainability to financial materiality that differ between companies. On the other hand, for small and mid-caps or unlisted companies, direct dialogue allows us to have a better understanding of the ESG practices of these companies. The ESG team meets with more than a hundred companies each year to discuss their sustainability efforts. On the other hand, the investment teams and the ESG team collaborate on engagement processes.

What resources do you have to do this analytical work?

The ESG research team is made up of six ESG analysts with multidisciplinary backgrounds that cover strategy, research and investment. The average team member experience is seven years in the industry. They liaise with management teams on ESG integration. Each ESG analyst covers specific sectors and analyzes the ESG quality of companies in order to better assess the risks and opportunities related to our investments. Most of the ESG analysis work is carried out internally for the European universes: of the 68 criteria that we analyze, 97% of this work is carried out exclusively internally. In addition, to perform this analytical work, we go directly to the source of the companies’ data. They also provide the management team with independent and comprehensive ESG research with detailed ESG recommendations and ratings of the issuers they cover. Alongside this, we also use external ESG rating and research databases or for biodiversity data.

This represents a gigantic mass of data. Do you carry out this analysis partly automatically?

ESG analysis not only requires processing a huge amount of data, but it often involves data that is not always standardized and comparable, as is the case with financial data. For certain environmental criteria – such as carbon, social or biodiversity emissions – the data available at the outset is often quite disparate in nature. This requires the use of sometimes very sophisticated processing techniques.

“Companies that best know how to integrate the social dimension of their activity are those that also perform best over time.”

Will you also use technologies such as artificial intelligence (AI) or machine learning?

We collaborate with a French company expert in data science and in the use of technology called natural language processing (NLP), a subset of artificial intelligence. With this technology, it is possible for us to perform semantic ESG analysis and more or less negative sentiments that can alert us to a major ESG risk linked to one or more companies. AI, and more specifically natural language processing, allows us to study the unstructured data that today makes up 80% of the world’s information available, including ESG data.

Isn’t there a risk that some companies will adapt the language used in their activity reports in order to escape this type of analysis or in order to appear in it in a more favorable light?

Of course, there is always a risk that a company will adjust its upstream data a bit to be perceived more favorably, or less unfavourably. This is why the use of AI can never replace the ESG analysis performed or at least supervised by humans. Technologies such as AI, machine learning or NLP are tools which are certainly very useful but which do not on their own replace the analytical work carried out by specialists who have the possibility of verifying whether the reality presented corresponds really to the reality of the company.

ODDO BHF also insists on the importance given to human and social capital in companies. Aren’t these aspects even more difficult to analyze and quantify than the environmental criteria?

Certain environmental risks – for example, CO emissions2 of an industrial company or the pollution caused to certain soils – can be directly measured and quantified by external companies or organisations.

When it comes to social criteria, the assessment methods are a little different and require looking in great detail at how a business operates. However, I believe as an asset manager that the social aspects, the “S”, can no longer be relegated to the background behind the “E” when analyzing a company. On the contrary, we observe that the companies that best know how to integrate the social dimension of their activity are those that also perform best over time. Despite this, the “S” has long remained the poor relation of ESG analysis – it is regrettable that some managers have often been content to use the services provided by external service providers. It’s not sufficient. In our analysis, we have always emphasized human capital, one of the two pillars of our approach alongside governance.

How do you broadly measure social criteria and what weight do you give them when evaluating a company from an ESG perspective?

Of the 68 criteria that we analyze in the ESG area, 15 of them are related to human capital. However, these criteria are weighted at 30% of a company’s total ESG score, which is a significant portion. ODDO BHF has a track record of more than 15 years in the analysis of data relating to the human and social capital of companies and our method of analysis has been able to be constantly refined in this area. It is a source of creation of added value which is essential in the medium and long term for companies.

We would love to say thanks to the writer of this write-up for this amazing content

AI will not replace human-supervised ESG analysis

You can view our social media profiles here and additional related pages here.