OPINION:

For years, ESG investing has been sold to retail savers, pension trustees and policymakers as a way to align portfolios with social values without sacrificing financial performance. Yet the data show otherwise. Portfolios built on ESG (“environmental, social, and governance”) ratings do not outperform conventional strategies — and may actively introduce risk.

The pitch for ESG has been remarkably effective: Global ESG-labeled assets surpassed $30 trillion by the mid-2020s, and the Securities and Exchange Commission, the European Union, and various state regulators have moved to mandate ESG-related disclosures on the theory that investors need this information to make sound decisions.

The empirical case for that theory is far weaker than the policy momentum suggests. In a new paper forthcoming in the Journal of Financial Stability, the authors ask: When ESG ratings are integrated into portfolio construction using standard quantitative methods, do investors actually do better? Using more than 22,000 firm-year observations from the Center for Research in Securities Prices and Refinitiv between 2003 and 2023, and running the analysis across 1,000 bootstrapped portfolios, the answer is no.

The argument for ESG integration rests partly on the claim that sustainability ratings capture forward-looking risks that conventional financial data miss. If that were true, portfolios that lean on ESG scores should be more stable, not less.

The data show the opposite. In portfolios of 200 stocks, ESG-only rules produce 6.3% higher annual volatility and 7.1% higher value-at-risk than return-based alternatives — meaning investors stand to lose substantially more in a bad year.

A one-standard-deviation increase in a portfolio’s ex-post ESG rating is associated with a 30-basis-point drop in mean returns and a 3.7% rise in volatility, both statistically significant at the 1% level. Whatever ESG ratings are doing, they are not improving the risk-return tradeoff.

The more revealing finding is what happens when investors do nothing special at all. Return-based portfolios — the kind any standard quantitative model would produce without consulting an ESG rating — end up with ex-post ESG scores statistically indistinguishable from portfolios built explicitly around those ratings.

A simple factor model that extracts the dominant component of market returns dominates ESG-based rules on both financial and sustainability metrics. The market portfolio itself achieves the highest ex-post ESG score of any strategy we test.

The implication is that the information ESG ratings claim to convey is already reflected in stock prices. Stock volatility predicts future ESG ratings, but ESG ratings do not predict future volatility. ESG scores are reacting to the market, not informing it.

This is not a minor technical caveat. Prior research by Florian Berg and coauthors has documented that ESG ratings from different agencies agree only modestly with one another — the correlation across major providers is roughly 0.5, compared with near-perfect agreement among credit rating agencies.

Refinitiv itself has retroactively revised historical ESG scores in ways that appear correlated with past stock performance. This raises the prospect that the ratings reward firms for being good investments rather than identifying which firms will be.

When investors use noisy, backward-looking ratings as inputs into portfolio optimization, the noise compounds. Estimation error in financial inputs already produces unstable portfolios; adding a noisy non-financial signal makes the problem worse, not better.

Our bootstrap analysis shows that ESG-integrated portfolios turn over significantly more frequently, which translates into higher transaction costs without compensating gains.

People are free to weigh environmental or social considerations as they wish, and the existence of ESG funds creates a legitimate option for those who want it. The work of Jeffers, Posenau, and Lyu, published in the Journal of Financial Economics, finds that impact funds underperform public markets by roughly 47 cents per dollar invested after adjusting for risk — a cost some investors will rationally accept in exchange for non-financial benefits.

The problem comes when regulators and fiduciaries treat ESG integration as financially prudent on the evidence currently available. Pension fund trustees who allocate to ESG strategies on the assumption of improved risk-adjusted returns are operating on a hypothesis that the data do not support.

Disclosure mandates premised on the idea that ESG ratings provide material, forward-looking risk information are mandating the production of data that markets already appear to discount.

A more honest framing would acknowledge that ESG investing is a preference, not an edge. Investors who want to express values through capital allocation can do so; those who want to maximize risk-adjusted returns are not obviously helped by integrating ESG scores into their portfolios.

Regulators should calibrate their disclosure regimes accordingly, focusing on the financial materiality standards that already govern securities law rather than treating ESG ratings as if they reliably capture information that markets are missing.

The case for ESG investing as a financial strategy needs to clear an evidentiary bar it has not yet met. Until it does, policy should stop pretending otherwise.

Stefan Padfield is a senior legal fellow at The Heritage Foundation and a principal in its Free Enterprise Initiative.