If you're an asset manager trying to boost returns today, chances are you have a sustainable investing strategy. A simple Google search for ESG — environmental, social and governance — yields some 44 million results, with the top hits touting headlines like, "The remarkable rise of ESG." Soon, the incorporation of ESG factors into investment strategies will become table stakes for these shops — 62% of asset managers held ESG funds as a quarter or more of their investments last year, up from 53% in 2017, according to a BNP Paribas study, and that number will likely only continue to rise.
However, this race toward ubiquity presents a problem: As ESG grows more prevalent, it's becoming harder and harder to generate alpha through these strategies. That's for a couple of reasons — first, everyone is using the same data. Eighty-six percent of companies in the S&P 500 now produce sustainability reports, up from just 20% in 2011, according to the Governance & Accountability Institute. That may seem like a positive development, and it is, but it also means that there is an influx of ESG data in the market that is being reused over and over again. All of this data gets rolled up by a handful of data providers and delivered through opaque ESG scores. That means that ESG is becoming more of a risk mitigation tool than an alpha generator. The second problem is that the majority of the data out there (the stuff that everyone is reusing) is being reported by the companies themselves. There are no universal, stringent standards as there are with financial statements, which means companies are effectively scoring their own businesses on sustainability. As a result, the market is left with very little useful data to employ in investment strategies.
So, what are asset managers to do? They've got to get creative. In some cases, that has meant collecting the data they need to make informed decisions themselves. In particular, companies are beginning to leverage unstructured data, such as news and social media, to uncover unique signals that have financial materiality using machine-learning techniques. In the future, there could be a greater focus on using actual data points to assess a company's sustainability, as opposed to relying on ratings from data providers.
Turning to alternative or unstructured data sources offers a path forward for asset managers looking to generate higher returns using ESG strategies. By looking at more unique data points, managers can find signals that may be overlooked by the masses; but also, it allows them to determine which factors they want to incorporate and which they don't. In short, pulling in thousands of raw data points in this way enables firms to determine what sustainability means to them. And new advancements in technology are making it possible to make use of more and more types of data. Machine-learning algorithms are now able to turn Glassdoor reviews into professional sentiment signals, for example, while monitoring satellite imagery can validate whether a company's claim to be in line with child labor laws is actually true. Data points are being collected all around us all the time, and the potential for this data to be turned into signals is almost limitless with the right tools.
In particular, the use of unstructured data can allow asset managers to identify specific ESG-related events and act on them before the market incorporates their impact. That's because investors are relatively inefficient at digesting ESG-related information, according to research from Deutsche Bank, meaning that those monitoring closely for the right signals in the news can get a leg up on their peers. For example, companies, on average, outperform their peers by 2 percentage points after the announcement of a litigation settlement, but it takes a full four months for that to be fully reflected in the stock price, according to Deutsche Bank. AI-based algorithms can flag events like this one, allowing managers to move early and find alpha. Perhaps one of the clearest examples of this is Facebook's stock performance following the Cambridge Analytica scandal. Word that the British firm had scrubbed data from 50 million Facebook profiles and used information it gleaned to direct the Trump presidential campaign in 2016 first broke in March 2018. However, Facebook's stock didn't drop materially until later that year when the company revealed that it was experiencing slowing user growth. The material ESG event was actually the initial announcement related to Cambridge Analytica; it just took a while for the potential financial impact to resonate with investors and trigger a drop in stock price. That means those that were able to identify the Cambridge Analytica news as financially material likely made off better than those who acted on the user numbers. The central idea here is that firms can use AI and advanced machine-learning techniques to determine which events are material before the broader market catches up, driving higher returns.
To be fair, there is still a long way to go before these techniques reach the mainstream. Not all asset managers have the ability to leverage machine learning just yet. And ESG's reporting problem isn't going anywhere fast. But the use cases that alternative data offer show that there is light at the end of the tunnel. As data science reaches more corners of the industry, and computing power generally increases, we are likely to see many more firms making use of machine learning to crunch vast amounts of data for a wide array of purposes, including to improve their ESG investing strategies. The data is already out there, ready and waiting; now, it's just about making sense of it.
Kate Drew is fintech research manager at Grant Thornton LLP, New York. This content represents the views of the author. It was submitted and edited under Pensions & Investments guidelines, but is not a product of P&I's editorial team.
