AI Governance in Impact Investing

ESG is high on the agenda for impact investors. Explore this thought-leadership interview with our guest Atul Monga, CFA to learn more about what this might mean for your organisation.

What is ESG or impact investing?

Investors around the world are increasingly seeking greener, ethical, and more socially responsible investment opportunities. Both governments and institutional investors want to be known for the real impact their investments are creating and don’t want to risk the fall-out from any bad publicity resulting from the slightest of linkages to investee companies that follow unacceptable practices. In today’s world, social media can create a significant impact on investor an customer sentiment and therefore corporate P&Ls. Recovery from the pandemic has in many instances also fast-tracked this mindset shift.

It should be said that business and strategy text books have long extolled the virtues of stake-holder management over pure shareholder management, but the benefits weren’t always quantifiable.

What practical areas does ESG cover?

ESG has many strands, the biggest of which is climate change impact – energy use, resource conservation, pollution, etc. But then there are other areas too like worker safety, treatment of animals, local community welfare, unethical market practices, political linkages, tobacco, and weapons, to list a few.

It would be fair to say that ESG investing can – to some extent – take the credit for forcing the reduced use of tobacco (in the western world, at least!).

Why are corporates increasingly focusing on ESG?

ESG investing has grown exponentially over the last few years – according to Bloomberg, global ESG assets are expected to grow to over $53 trn., more than a third of the total $140.5 trn. assets under management globally. And that’s just equity alone. Then there is debt, which could also grow from $2.2 trn to $11 trn in AUM by 2025 too. By focusing on ESG, a corporate stands to gain significantly not only through being more sought after as an investment but also from generating better customer affinity. And, if a true link could be established between a greater focus on ESG and enhanced financial performance, that would be an added bonus.

Why is it such a complex area?

Firstly, one can’t simply rely on what companies claim or self-report, either by picking and choosing selectively what they report, or just focusing on the cause rather than the effect, or simply the fact that information is not all quantitative and therefore hard to make sense of. While some ESG factors are easy to deal with by using exclusion-based screening of investment opportunities, others aren’t simply a case of black and white. For example, it is easy to exclude companies that have any linkage to the promotion of tobacco or weapons or rely heavily on hydrocarbons. But areas like climate change, worker safety, treatment of animals, etc. require complex measurement rather than exclusionary assessment. To make life more difficult, disclosed data is ridden with biases, so measuring real impact requires dealing with structured and unstructured data.

So how does AI help?

Just as algorithmic trading has made the complex world of quantitative investing easy, algorithms can play a big role in impact measurement and establishing correlations, in the complex area of ESG. Dealing with vast amounts of data – structured and unstructured, across multiple sources, combining multiple areas of ESG – and differentiating cause and effect into ratings that are easy to understand is not humanly possible. At best, humans could use sampling to deal with the volumes of data involved but then as behavioural finance has shown us, biases can set it, impacting the value of such analyses.

This is where AI and machine learning can help with ingesting vast sums of data to establish measurable linkages with actual performance.

Is AI the solution then?

Well, yes and no! AI can deal with the challenges of ESG data volumes, but what it can’t deal with is the limitations of the data itself. Whether it is the fact that we can only rely on known sources of data at any point in time or the underlying correlations that known sources of data might already represent, AI itself is prone to biases. This is where governance of AI models is important. Mitigating these biases through constant monitoring for feedback loops and adding in new relevant sources of data will help enhance the outcome and make for better investing and purchase decisions for the world at large.

Does the finance community have any other interest in AI?

Yes, indeed. As you can imagine, AI is a hot investment area in its own right. RPA, ML and now AI businesses have been closely watched by the investment markets for years. Investors are excited about companies using AI themselves and understanding whether they are doing so effectively and ethically to ensure they don’t fall foul of the ESG agenda themselves. What is the impact of using AI on a company’s workforce composition? Is it creating efficiencies? Any biases? How does society view it? And then of course there are AI companies themselves and whether they themselves qualify for ESG investing, through advocating ethical use of AI in the first place. But this is a whole new discussion topic in its own right.