There's more information than ever on ESG investing, but it isn't helping

Mint

08-07-2025

Defenders of environmental, social and governance investing have mostly given up their strongest claim, that of doing well by doing good: That you can beat the market while helping improve the world. A new study finds that even their weaker financial claims don't stand up to scrutiny.
The fallback position for defenders of environmental, social and governance investing in the face of political and legislative challenges in the U.S. is typically that ESG should be considered in their stock picking as financially material information, that it can help make a portfolio more stable and that anyway it is better to have more information than less.
What BlackRock and other big fund managers call 'ESG integration" involves taking information such as corporate carbon emissions, labor relations or human-rights infringements into account when deciding whether to buy or sell a stock.
The obvious problem with this approach is one I've discussed before: In theory at least, a stock with a terrible ESG record could still be a screaming buy if it were cheap enough.
Research by French university spinout Scientific Beta, part of SGX Group, confirms the difficulty. They constructed a portfolio to make optimal use of 15 years of ESG information, balancing risk and reward while maintaining diversification using standard tools.
In many cases it would have taken the anti-ESG position, buying stocks such as tobacco producers, buying companies generating environmental controversies and selling companies that spend a lot on employee training or have strong employee involvement in the community.
For a purely financial investor this wouldn't be a problem. If you just want to make money, then of course coal miners or oil producers will be worth buying at some price, even if their products are highly regulated, face special carbon taxes or make it harder for them to recruit young workers. But anyone who called this ESG investing would be laughed at and potentially face legal challenges.
The next defense is that ESG information can be used to make a portfolio less volatile. But it turns out it makes no difference. The new research by Scientific Beta's Giovanni Bruno, Felix Goltz and Antoine Naly looked at more than 200 ESG factors used to optimize a portfolio to balance performance and risk, and found that they offered no improvement over traditional financial factors.
'There's just not evidence of an incremental return contribution from these ESG metrics," Goltz said. 'If you have traditional financial objectives you don't really need these ESG metrics."
That doesn't shoot down the core aim of many investors who object to their money being used to finance industries or issues they oppose. But they should give up on the idea that what is good for the planet will also help them beat the market.
Here we come back to the theoretical problem: What is priced in. Even if ESG is good for the bottom line, if that's already priced into the stock then there's no reason the shares should do well in the future.
A politically uncontroversial, if surprising, example of this: Companies with strong audit committees would be sold when the portfolio was optimized, Goltz said. My assumption is that they do relatively less well because it's so well-known that audit quality matters that investors put too much weight on the issue, and it is overpriced.
What about the idea that more information is always better? Well, it turns out a lot of information doubles up with standard financial metrics—perhaps because more profitable companies have the capacity to spend time and money improving their ESG ratings, rather than because better corporate ESG improves profitability. If extra information adds nothing new, it is simply a costly distraction.
Even for the many ESG issues that do seem to provide valuable information in backtests, better performance mostly disappeared when used for forward-looking investment. To test this the researchers divide up the 15 years into three periods, optimize portfolios using data from two of the periods and see how it did in the last period. Looked at 'out of sample," as statisticians say, the ESG metrics overall added no value.
Such problems aren't unique to ESG. Financial researchers spent decades identifying factors such as accruals, earnings smoothness or stock liquidity that beat the market, only to find that almost all of them stopped working as soon as the work was published.
Some of this might be because prices adjusted to take them into account, but some was surely cherry picking; over any period there will be some features that worked better than others, but that doesn't mean they were the cause or will keep working.
The easiest way to make money from ESG might be to cancel the subscription to ESG data.
Write to James Mackintosh at [e-mail]