ESG data doesn’t help investors make better risk-return decisions – study

Using ESG metrics does not build better portfolios than if investors just used traditional financial data, according to new research carried out by Scientific Beta.

The index solutions provider found that there are financial performance benefits from integrating ESG information, but that this added value disappears when an out-of-sample analysis is conducted.

This is when portfolios are built only using the information that is available at that point in time and assessing the results on that basis.

Its analysis also uncovered that a financially optimal use of ESG information would involve some “anti-ESG investing”, meaning tilts emphasising companies lagging on certain ESG factors.

Third, the researchers found that in an out-of-sample analysis, adding ESG characteristics does not improve risk mitigation.

To carry out its research, the firm built an ESG dataset of 222 characteristics from various sources, including both traditional ESG ratings based on analyst opinions, and alternative ESG characteristics that rely on systematic textual analysis of news or firm documents.

The standard financial dataset reflected 130 characteristics measuring valuation, momentum, quality and similar financial ratios.

In conducting its analysis, Scientific Beta’s approach was deliberately focused on the perspective of investors who care only about risk-adjusted returns, framing ESG information as an input only.

The “clear message” for investors, according to Scientific Beta, is that “they need to set realistic expectations about what ESG integration can achieve”.

The lesson extends beyond ESG, it added: “Any claim of informational advantage must survive out-of-sample testing, not just traditional backtests.”

Felix Goltz, research director at Scientific Beta, told IPE the research also suggested that investors need to be more frank about the potential trade-offs involved in ESG integration, namely that pursuing optimal financial objectives could come at the expense of non-financial goals.

As regards out-of-sample analysis, he said this research method provides a more realistic framework, with traditional backtesting carrying the benefit of hindsight.

He said that out-of-sample analysis was by now fairly common when it came to topics like traditional factor investing, but that it was rarely applied to ESG investing.

“I myself am guilty of having written lots of papers with in-sample analysis but everyone should look more at out-of-sample analysis,” he added.

The latest digital edition of IPE’s magazine is now available