Information about a company’s environmental, social, and governance (ESG) behaviour may reveal how risky a security is on a statistical basis, according to investment manager AQR.
In AQR Capital Management’s new paper – “Assessing Risk through Environmental, Social and Governance Exposures” – the firm said it found a strong positive relationship between companies’ ESG exposures and the statistical risk of their equity.
“Stocks with poor ESG exposures tend to have higher total and specific risk and higher betas, both contemporaneously and as far as five years into the future,” the Greenwich, Connecticut-based manager said.
“We interpret these findings as evidence that ESG information may play a role in investment portfolios that goes beyond the ethical considerations and may inform investors about the riskiness of the securities in a way that is complementary to what is captured by traditional statistical risk models,” wrote the paper’s authors, Jeff Dunn, Shaun Fitzgibbons, and Lukasz Pomorski.
Investors who wanted to tilt their portfolios towards safer stocks might be able to combine the two to build more stable and robust portfolios, they said.
AQR said it found clear support for its hypothesis that ESG exposures could be informative about the risks of individual firms in data which it said was robust to a wide variety of controls and various stock universes.
“We also find that ESG scores may help forecast future changes to risk estimates from a traditional risk model,” AQR said.
Poor ESG exposures predicted increased future statistical risks for equities, the authors found.
However, when broken down into the three components of ESG risk, AQR’s research team found that it was the social and governance pillars that showed the strongest correlation to risk.
“The environmental pillar is only insignificantly related to the various risk measures,” the authors said.
“On the one hand, it may be that environmental exposures are inherently less predictive of companies’ risks,” Dunn, Fitzgibbons, and Pomorski wrote. However, it was also possible that data on environmental exposures was “noisier” than data on the other two components.
This noise in the variable could be preventing computations from giving more precise, statistically significant estimates, the authors suggested.
The full paper is available here.