Is artificial intelligence (AI) the future of investment decision making? Crimson Black Capital has staked its fortunes on that idea. Launched in 2019 and unburdened by legacy systems or thinking, the London-based fund firm went all-in with AI from inception. Its proprietary data-mining and portfolio construction engine leverages multiple AI-powered data sources and uses machine learning algorithms to generate long-short portfolios, with the goal of consistently outperforming the markets.
“We are completely AI-driven,” says co-founder and Chief Investment Officer Ayush Ansal. “We apply a variety of AI models to millions of inputs from traditional and alternative data sources. Those models assess the inputs, rather than human beings, and generate outputs that inform our investment decisions and trading.”
ESG factors figure prominently in the inputs that drive the firm’s security selections. And for that, Crimson Black chose Truvalue Labs, a FactSet company.
“We use Truvalue Labs’ signals as a core part of our investment process,” Ansal explains. “We combine those signals with other proprietary signals to develop our trade baskets. We’re very focused on ESG and how to implement it in a way that is alpha-accretive, rather than just looking to be green.”
Indeed, Crimson Black is not marketing its funds as “green” or socially responsible. Rather, Ansal and his team have identified a strong correlation between favorable ESG performance, as measured by Truvalue Labs, and the performance of a stock. “What we’ve noticed is that Truvalue Labs is often selecting the same equities as other AI models in terms of alpha indicators,” Ansal points out. “We are genuinely using the Truvalue Labs’ ESG data to improve our alpha.”
Taking Sentiment Analysis to a New Level
Truvalue Labs uses AI to scour enormous volumes of unstructured data sources to learn what is being reported about companies in near-real time that is indicative of ESG behavior. This is in sharp contrast to widely used, conventional ESG data and ranking sources that rely largely on self-reported company information at annual or semi-annual reporting intervals. Before selecting Truvalue Labs for ESG data, Ansal had conversations with some of those traditional providers. “The reporting from other ESG sources is based on a company’s own disclosures and very limited independent analysis,” he says. “That is not really going to help us generate alpha. That kind of ESG data may be applicable to green funds, but there’s no way we can beat the S&P 500 using it. If we want to actually generate alpha, Truvalue Labs data is the way to go.”
The results Crimson Black has seen in its portfolios have convinced the firm it made the right decision. “We have run our portfolios [through Truvalue Labs] for analysis for several months, and it’s been a definite value-add,” says Ansal. “The way that Truvalue Labs’ ESG data is useful is that it covers a very different type of news source. The truth is that the market is extremely sentiment-driven. Quickly identifying and incorporating company-specific local news data is very useful in identifying different types of alpha that other AI models can’t.”
Early Warnings of Risk Ahead
Truvalue Labs is designed to help uncover both risks and opportunities that aren’t readily apparent from other sources. The platform generates several scores, including an Insight score, ranking a company on ESG behavior at any given moment, plus a Momentum score, showing the company’s trajectory on ESG perceptions, upward or downward, over time. A potential opportunity arises, for example, if a company has brought on new management that promises to correct prior behavior.
“Truvalue Labs’ data is very helpful as it identifies relative changes,” Ansal says. “It won’t just tell me that Company A is good and Company B is bad. It will tell me Company A is getting worse and Company B is getting better. That change is what produces alpha, rather than just telling me what I already know.”
Truvalue Labs also provides early warning signals on companies that may have a rough road ahead. “Where ESG analysis is very useful is in preventing losses,” Ansal says. “Before any other model, for instance, Truvalue Labs’ ESG data will tell me if a social media company is having too many problems in the media with privacy, or an aircraft manufacturer is having problems with operations. In the world of AI-based investing, where transparency is often sacrificed for higher returns, this platform allows me to drill into specific events that are guiding the model’s decisions and to learn why certain equities are being avoided or shorted by the system. Truvalue Labs can give me that warning before any other source, because these events play out in the news a lot more quickly than they will show up on earnings. That’s a huge differentiator compared to other AI models.”
Understanding the Sources of ESG Performance
The platform further allows users to drill down into actual data sources that shed additional light on the reason for a company’s scores. “With Truvalue Labs, it’s very easy to see the Spotlight Events™ that are most relevant to a particular company, and then link to the underlying news articles. That’s amazing. I can’t tell you the value that adds to our investment process. Even though we are quantitative, our investors are human and so are we. We do like to read about companies we are following. It’s very useful to be able to zero in on the most relevant news.”
Crimson Black has already seen success in generating returns in line with expectations, as well as in attracting investors to its novel approach. “Investing is fairly simple,” Ansal says. “You get data points about the market, you analyze those data points, and you trade based on that analysis. AI is the newest and most cutting-edge way to do that analysis. We wanted to incorporate as many different types of inputs as possible, using machine learning to give us the most useful predictions to select our equities. Truvalue Labs’ ESG data is different from everyone else’s, and very clearly brings a valuable perspective to our process.”