One label covers different objectives
Environmental, social and governance information can be used to estimate financial risk, express preferences about holdings, or seek changes in corporate conduct. Those purposes are related but not identical. The SEC’s investor bulletin describes funds that integrate ESG alongside conventional financial inputs, exclude certain investments, or engage with companies to improve practices. It also warns that similarly described funds can follow different approaches. [1]
An exclusion strategy may deliberately avoid a profitable company because of the activity it conducts. A financially focused strategy could own the same company if the valuation adequately compensates for its risks. An engagement strategy might hold it precisely because management has room to change. None of those positions can be evaluated coherently until the objective is clear. A portfolio that meets one mandate can disappoint another without either description necessarily being fraudulent.
Ratings disagree about more than weights
Berg, Kölbel and Rigobon’s 2022 study compares six ESG ratings using 709 underlying indicators. Their decomposition attributes 56% of ratings divergence to measurement, 38% to scope and 6% to weighting. These percentages describe their sample and methodology; they are not universal error rates for every provider or every subsequent rating. [2]
Scope concerns what is included, such as lobbying or workplace practices. Measurement concerns how a particular attribute is assessed. Weight concerns how much an attribute contributes to the aggregate. An apparently simple disagreement over which company is better can therefore combine different questions, different observations and different aggregation rules. A composite grade can also hide offsetting environmental, social and governance results. Disagreement is evidence of measurement complexity, not proof that every underlying datapoint is useless.
High past returns can coexist with lower expected returns
Pástor, Stambaugh and Taylor’s sustainable-investing model distinguishes the return investors expect when buying an asset from the return produced by unexpected changes in preferences or climate-related concerns. In their framework, greener assets can have lower expected returns because investors value holding them or their hedging characteristics, yet outperform when demand for greenness rises unexpectedly. This is a theoretical equilibrium result, not a forecast for a particular fund. [3]
The distinction resolves an apparent contradiction. A rising valuation creates a gain for existing holders while leaving a smaller prospective return for a new buyer if future cash flows are unchanged. Conversely, an ESG characteristic could be associated with valuable operating resilience that the market had underestimated. Observed outperformance alone cannot distinguish repricing, cash-flow improvement, sector exposure or luck. The investment label does not identify the mechanism.
Assets with an ESG policy are not all an ESG-driven allocation
In Green Tilts, Pástor, Stambaugh and Taylor estimate portfolio tilts associated with ESG characteristics rather than simply adding up all assets managed by institutions that state an ESG policy. The paper’s central measurement distinction is between an institution’s whole portfolio and the deviation from a counterfactual allocation without ESG preferences. The counterfactual is estimated, not directly observed. [4]
Consider a purely hypothetical $100 million fund. If a particular screen changes only $5 million of holdings relative to an otherwise identical strategy, calling all $100 million ESG assets answers a different question from measuring the $5 million allocation difference. The example does not imply that the remaining assets are unaffected by engagement or risk analysis. It shows why asset totals vary with definitions and why adoption of a policy is not itself a dollar measure of additional investment.
Portfolio greenness and real-world impact are separate outcomes
Analytical illustration: selling shares in a high-emissions company immediately changes the seller’s portfolio footprint, but another investor owns those shares afterward. The trade does not mechanically close a factory. A real-economy effect could instead arise through financing costs, new capital provision, corporate engagement, customer demand or policy. Each channel requires evidence beyond the portfolio’s before-and-after rating.
Allcott, Montanari, Ozaltun and Tan develop an economic measure of corporate social impact based on the welfare loss from a firm’s exit. Their study covers 74 firms across twelve industries, using survey evidence and models of substitution in product and labor markets. Existing ESG and impact ratings were essentially unrelated to their proposed measures. This is an important competing framework, not a universally accepted definition of social welfare. [5]
The counterfactual matters: replacement suppliers, workers’ alternatives and external effects can change the outcome substantially. A company’s activity, a shareholder’s contribution to changing it and a fund’s reported footprint are three different units of analysis. Measuring one cannot automatically establish the others.
Disclosure policy is a dated legal question
The SEC’s official rulemaking page records withdrawal, effective June 17, 2025, of the 2022 proposal for enhanced ESG-practice disclosures by certain advisers and investment companies. That proposal should not be described as an operative disclosure requirement. This status was checked against the SEC page on October 4, 2026. [6]
That narrow procedural fact is not a conclusion that every ESG-related obligation disappeared, or that rules in other jurisdictions have the same status. It also does not settle separate questions about issuer climate disclosures, pension duties, advertising or misstatements. Those regimes involve different instruments and authorities. This article makes no general claim of regulatory approval or prohibition of ESG investing.
What a performance claim needs to mean
Analysis: a financial comparison depends on a specified benchmark, period, currency, fees, risk exposures and investable information set. Sector exclusions can alter returns even if the excluded characteristic has no independent predictive power. Using ratings revised after the investment date introduces hindsight; ignoring closed funds can introduce survivor bias. Correlation between a rating and profitability is not enough to show that the rating caused the profitability.
The same discipline applies to impact claims. A change in a portfolio metric is useful evidence about holdings; a change in company behavior is evidence about that company; attributing the latter to an investor requires an additional causal argument. ESG is therefore better understood as a collection of mandates and measurement choices than as a single asset class with one promised financial or social outcome.
Sources
- SEC Investor.gov, Environmental, Social and Governance Funds — Investor BulletinOfficial sourceBack to text: ↑
- Berg, Kölbel and Rigobon, Aggregate Confusion: The Divergence of ESG Ratings; Review of Finance 2022SourceBack to text: ↑
- Pástor, Stambaugh and Taylor, Sustainable Investing in Equilibrium; NBER 26549, revised June 2020, published 2021SourceBack to text: ↑
- Pástor, Stambaugh and Taylor, Green tilts; authors’ account of their original research, CEPR, August 13, 2023SourceBack to text: ↑
- Allcott, Montanari, Ozaltun and Tan, An Economic View of Corporate Social Impact; NBER 31803, April 2025 revisionSourceBack to text: ↑
- SEC, withdrawal of ESG investment-practice disclosure proposal, S7-17-22; June 2025Filing / reportBack to text: ↑