Stuart Doole heads the development of new indexes at MSCI, where he leads a global team that builds indexes tailored to the needs of a range of institutional investors. In the following Q&A, Stuart discusses his conversations with investors since the start of the coronavirus pandemic, the evolution of ESG investing since the launch of the MSCI KLD 400 Social Index and how MSCI uses AI and machine learning in developing ESG indexes.
In your recent conversations with investors, what are you hearing?
The coronavirus pandemic has investors looking at the sustainability of companies in a broader sense. The sense is that ESG is supposed to assess companies’ resilience to exogenous events, as our research suggests that it has. Apart from the pandemic, investors are interested in how they can build standards into indexes that allow them to, say, distribute a sustainable exchange-traded fund across countries that each have their own set of rules. Some central bankers with whom we’ve spoken are interested in how ESG ratings map to the UN Sustainable Development Goals (SDGs), which can inform their view of how micro interests build up to macroprudential interests.
Can you talk a bit about building ESG indexes?
Our job is to anticipate how investors may apply ESG investing to strategies, and what benchmarks or indexes they may then need. For example, with factors, we saw some early interest from clients in strategies that might embed ESG into a minimum-volatility approach and also in ways to reduce the environmental exposure of value investing. So we built a framework for incorporating ESG into factor indexes. A few years down the road, strategies involving ESG and factor indexes have gathered a lot of assets. Investors want indexes tailored to their unique needs, whether to reflect their own views or because it’s something their investment committee has directed. So, for example, if a pension fund supports efforts to limit the rise in global temperatures by 2 degrees Celsius through its investment strategy, it may want the carbon footprint of its index to be, say, 30% lower than the market-capitalized index alongside a reduction compared to prior years and improved systematic allocations toward climate-solutions companies. It’s very natural that we customize indexes to meet client needs but you can only do that in a responsive and nimble way if you’ve done the background work. With our knowledge and our tools, we can do that.
Where is ESG index construction going, particularly with advances in artificial intelligence and machine-learning applications?
The inputs for ESG ratings have always been a mixture of corporate disclosures and alternative data, such as information from third parties or nongovernmental organizations. With artificial intelligence, we’re able to train computers to help identify themes for indexes as well as risks for companies. We can leverage machine-learning techniques such as topic modeling and natural language processing to inform our data gathering for ESG datapoints, as well as create flexibility and scalability for ESG indexes. For a thematic index like clean tech, you might start with a concept and a description or some key ideas. But you’re still a long way from building an index of individual companies. We don’t go out and search the internet for a document that happens to mention, for example, Apple and clean tech. That would lead you to all sorts of nonsense. Instead, we might build a dictionary of words that are dense in documents, whose relevance to our concept we can quantify and that we can match up against companies’ business filings, websites and other information to help us identify possible constituents.
As the work becomes more automated, how is the role of ESG analysts changing?
With automation, the volume and range of information available keep increasing. While the machines provide a first level of data selection and analysis, we refine it with the help of analysts, which raises the quality and richness of our analysis significantly. Remember that everyone used to do all this work manually, so we were at a high quality already. The technology improves our ability to analyze and extract meaning from unstructured data on a much larger scale. So we can investigate many more companies more quickly. By using machines to preselect information, we allow analysts to focus on their industry expertise. The machine does a first review. The analyst has the final say. It’s a marriage of the two, but there’s always a human in the mix.
How has ESG integration evolved in recent years?
When I joined MSCI, there was much less demand for ESG index products. To the extent that people wanted to customize, most of it focused on values-based exclusions. The pickup in customization came as investors wanted more nuanced exclusions. For example, for an index that excludes tobacco stocks, people wanted the exclusions based on the precise business-model exposure of individual companies. The reduction in carbon footprint or the sustainability profile of an index is now part of almost every conversation, compared with six years ago, when things were in very neat boxes and we had far fewer hybrid products. Today, clients who want us to develop indexes around such things as demographic trends or robotics or electric vehicles also want some ESG screening factored into the final index design. The same is true for asset owners wanting us to build multi-factor indexes for them.
How do you think about the value of the MSCI KLD 400 Social Index as it begins its fourth decade?
It’s always fantastic to be able to show something that has been running for so long with both values-based exclusions and a more rounded assessment of a company. And it’s been live in the U.S. as well, which has been a more difficult market for ESG investing ideas until recently. While the methodology and some of the key inputs have changed over time, the index has remained true to trying to hit its objective of exposure to companies with outstanding ESG ratings while excluding those whose products have negative social or environmental impacts. So the MSCI KLD 400 Social Index really does offer us a window on how ESG has played out over 30 years as an index-based approach.
What makes ESG challenging as a strategy?
Our research shows how ESG considerations have been financially relevant in a broad sense. Still, ESG can be challenging for a lot of investors and portfolio managers, who are not yet used to applying these considerations systematically or comprehensively to their portfolios. Some have embraced the ideas behind the “E” and the “G,” but the pandemic has illustrated yet again that issues like the supply chain, human capital, corruption and how companies treat their workers can pose significant risks to both active and passive strategies. It all goes to the resilience of businesses.
What are you seeing in terms of each of the three ESG pillars as it relates to index construction?
It depends on your time horizon and the industry. But in general, when we pull apart the pillars, we find that over a short period of time like a year, governance has proved to be more relevant than environmental and social indicators in terms of their impact. By contrast, environmental and social indicators have tended to play out over longer periods.
As a company, MSCI is committed to supporting sustainable investing. That’s why we published “The MSCI Principles of Sustainable Investing,” which set forth MSCI’s views and recommendations on the core principles and best practices for ESG integration by investors globally.
Anecdotally, with clients, we are currently seeing see a lot of focus on the environmental pillar because of climate change. There’s always interest in governance as a separate pillar because that is where ESG is seen to be the most quantitative, with the longest track record of data, and as a risk-management overlay to another index strategy. Interest in social issues was appearing more often lately, even before the COVID-19 crisis. It comes in through index exclusions and also because more clients are interested in impact investing and the UN SDGs, which are increasingly used to express investing objectives and define reporting to stakeholders, especially related to climate, diversity and human capital. It really depends on each investor’s objectives.