The Unintended Consequences of ESG

A measurement system does more than describe the world; once money follows it, the system helps change the world it measures. ESG scores can push companies to disclose risks and improve practices, while also encouraging attention to whatever a scoring method can count most easily.

Suppose a school wants students to understand science and begins rewarding the number of lab reports submitted. The count may encourage practice, but students could learn to optimize report volume rather than understanding. ESG ratings face a similar design problem: once a proxy affects funding, organizations may improve the proxy even when it only partly represents the intended outcome.

The label covers different questions. One investor may care about how environmental and social issues affect a company's financial risk; another may want capital to support particular outcomes. Ratings providers can differ in scope, weighting, and data. Two scores bearing the same label may therefore answer different questions rather than contradict one another.

Imagine a company deciding between reducing a hard-to-measure supply-chain harm and improving a well-publicized disclosure metric. If investors reward the second more visibly, managers may prioritize it even when the first matters more to affected workers. The incentive need not be dishonest: people naturally allocate scarce time toward goals that are legible and rewarded.

Capital allocation can shift as ratings enter investment screens, lending terms, procurement, and executive targets. A lower score may raise financing costs or exclude a firm from a portfolio. That pressure can motivate change, but it can also move capital away from a difficult sector without improving the underlying activity, depending on what the investors and firms do next.

There is a counterintuitive risk in rewarding the cleanest-looking companies: firms with large transition challenges may receive less support to make improvements. Divestment can express a boundary, yet engagement or transition finance may sometimes influence real operations more directly. Which approach works depends on leverage, transparency, and credible milestones.

Consider two investors: one excludes a high-emission company, while another finances a verified plan to reduce its emissions. The first can avoid exposure; the second may influence operations, but only if the plan is credible and the investor has leverage. Neither method is automatically superior. The relevant question is what changed, for whom, and whether the change would have happened anyway.

ESG systems also have limits that should not be confused with proof of fraud. Data gaps, inconsistent definitions, and the difficulty of comparing different industries can make a single rank falsely precise. A utility and a software firm face different impacts; a common score may conceal more than it reveals unless its assumptions are clear.

The strongest opposing case is that imperfect scores are still useful. Without comparable disclosures, investors may rely on even less information, and a public framework can expose risks previously ignored. The answer is not to demand a flawless number but to treat ratings as evidence with a method, a purpose, and uncertainty attached.

Companies can respond more responsibly by reporting outcomes, explaining trade-offs, and distinguishing near-term targets from verified changes. Investors can state whether they are managing financial risk, seeking impact, or doing both. Clear objectives make it harder for a convenient metric to stand in for the actual goal.

Regulators and standards bodies can improve comparability while guarding against false certainty. Auditable definitions, consistent materiality rules, and disclosure of rating methodology help users see what a score omits. Better measurement will not settle moral disagreements, but it can make them harder to hide inside a composite number.

ESG is most useful when it opens a conversation about consequences rather than closing one with a grade.

What evidence would convince you that a responsibility score is changing outcomes rather than merely changing how an organization presents itself?