Investors Turn to Raw Data Over Ratings in ESG Alpha Hunt

Product returns are not usually a factor in socially responsible investing. But with more conventional approaches that rely on environmental, social and governance (ESG) scores looking increasingly dated, some quant firms are combing novel datasets, such as textual information on product returns, for more granular investing signals.    

Technologies such as natural language processing can be used to extract meaning from text documents, making it possible to identify and test previously unseen signals, said Peg DiOrio, head of quantitative equity portfolio management at Voya Investment Management. Information on product returns can be found in accounting disclosures, for example.

“When a company has junky products, that are defective, that don’t meet customers’ demands, that are likely to be returned – that also destroys your brand,” said DiOrio, adding that high-quality products command more pricing power.

DiOrio was speaking on a panel at Risk.net’s Quant Summit Virtual conference on July 14.

Speaking on the same panel, Riti Samanta, head of fixed income strategy at GMO, said her firm ran an emerging markets fixed income strategy that takes account of governance. GMO uses raw data including text data to conduct this analysis, she said, rather than rely on off-the-shelf scores.

A minimum amount of data is required to get a quant strategy up and running, Samanta said. “That naturally leads towards using the larger [data] providers to get a consistent dataset to start from. But that’s really the starting point.”

ESG scores from providers such as MSCI and Sustainalytics are comparable to credit ratings and play an important in ESG investing – but they are not a source of alpha, panellists said. “Something that’s rated triple-A: that’s a good start but it’s just a starting point,” DiOrio told the audience. “You need to dig. You can’t just pull out a chequebook buy some canned data and call it a day.”

ESG investing is growing fast. According to analysts at Bank of America, the sector is set to grow by $20 trillion in the next 20 years. BlackRock has said it will integrate ESG in all its active portfolios by the end of the year.

At the same time, the poor quality of data histories and a lack of comparability between ESG scores from major vendors have been gripes among quants trying to add an ESG element to their investing.

Leading providers of ESG ratings such as MSCI and Sustainalytics concur on only about half the companies they rate, according to data from State Street Global Advisors. Quants typically backtest strategies over periods of 10 years or more, but the data is patchy and seldom covers that length of time.

Social metrics have in the past been seen as the least reliable element of the ESG composite. But investors see things like employee welfare as an increasingly meaningful component in driving company profitability. A survey from BNP Paribas Asset Management found a 20% jump in the number of institutional investors that expect social metrics to be “extremely” or “very” important in investing decisions after the pandemic.

Newfound signals relating to staff welfare, for example, have proved their value during the Covid-19 crisis, DiOrio said. “A lot of your employees getting sick, your facility getting closed, [being] in the headlines – that’s bad for your brand, bad for your stock price.”

Fashion firm Boohoo’s stock price has nearly halved after the company was accused of buying from garment workshops where workers were underpaid and poorly protected from Covid-19 infection. 

“These are niche factors. This is not necessarily the ESG that people talk about,” said DiOrio. “The data that is most important is unstructured,” she added. “It’s more nuanced. It’s more specific.”

Arabesque S-Ray sells raw and aggregated data to investment firms. Todd Bridges, the firm’s head of sustainable investing and ESG research, said the most sophisticated quant hedge funds were interested mainly in raw data and are developing their own signal weightings and rating schemes in-house.

His advice to quants wanting to take advantage: “You’re going to have to get your hands dirty and really understand the intricacies of the data.”