By 2 November, ESG rating providers that want to keep operating in the EU must file with the European Securities and Markets Authority (ESMA).
The shorthand for what happens next is that ESG ratings are now regulated. That is true, and mostly for the better.
ESG ratings have been sold into investment processes for two decades with no dedicated EU supervision. Separating the ratings business from the sale of services to rated entities removes the most corrosive conflict of interest in the market. Concentrating supervision in one authority rather than 27 avoids fragmentation.
The Annex III disclosures do something new. An investor will finally be able to see whether two providers disagree because they used different data, or because they weighted the same data differently.
On methodology itself, the regulation leaves providers free. That is the right call. Much of what looks like methodological divergence is judgment, and judgment is what investors are interested in. Pushing two providers towards the same answer leaves investors paying twice for a single opinion.
Where standards slip
Quality, then, rests on a single assumption. Once conflicts are removed and methodologies are published, competition is expected to look after the rest.
The green bond market has already tested that assumption, and it is the nearest neighbour an ESG rating has. A verifier checks a bond against a label, while a rater grades risk, impact or both, so the two products differ. The economics underneath them match. Both sell expert judgment that buyers take on trust, and both draw credibility from the reputation of the field as a whole.
According to a model of green bond certification published this year, that shared reputation is what erodes standards. Being strict is expensive for the certifier that does the checking, while the credibility spreads across every rival because investors judge certifiers as a group. Each firm carries the full cost of rigour and captures only a fraction of the reward. Each firm therefore relaxes its standard, and competition accelerates the decline.
The numbers are clear. Two competing certifiers settle at just over half the standard a single certifier would maintain. Five competitors settle at a quarter of it. A lower standard means the label extends to weaker projects.
The certifiers in this story are honest. The model gives them full independence and zero conflicts of interest. Yet standards still slide because shared reputation alone is enough. The conflict-of-interest rules in the regulation solve a real problem. The erosion described here is a different problem, and it survives them.
ESMA’s notification list, published in July, already names Sustainalytics, MSCI, Sustainable Fitch, Clarity AI and EthiFinance, among others. That is a crowded field and, in this setting, crowding is itself the source of pressure.
A scoreboard now, a floor next
The model points to two remedies. Both leave methodology in providers’ hands, and one is already within ESMA’s reach.
That one is a public record of performance. In the model, standards climb when investors learn faster who was right. Put simply, certifiers are stricter when someone is keeping score.
ESMA already maintains exactly such a record. CEREP, its public repository of credit rating performance statistics, has been updated twice a year for more than a decade. The credit rating rulebook makes the same promise of keeping supervisors away from methodologies, and CEREP has operated comfortably alongside it.
An ESG scoreboard would need its own yardstick, and the regulation has supplied one. Credit ratings are judged against defaults. ESG ratings measure risk, impact or both, and the new rules make each provider say which one.
That statement is the benchmark. ESMA can hold every product to its own stated objective, publishing rating histories, revision frequency, data vintages and the share of each score built on estimates instead of reported figures.
The infrastructure is arriving on schedule. ESAP, the EU’s new central data repository, began receiving corporate sustainability disclosures in July, and data filed under the ESG Ratings Regulation joins it from January 2028. Turning that library into a scoreboard would take a set of statistics and perhaps some additional reporting rules, still well short of reopening the regulation.
The floor is the more demanding reform. It would sit under what a rating may rest on, setting how much of a score may be built on estimates and how fresh the underlying data has to be, while weightings stay free. Providers would keep the freedom that matters: the freedom to disagree about what the inputs mean.
Yet the regulation promises to keep ESMA away from rating content and methodologies, and a binding rule on inputs presses against that promise. It probably needs a change to the Level 1 text.
The model supplies the reason to persist. The more crowded the market becomes, the higher the floor needs to sit, and this market is growing more crowded. The European Commission’s review of the regulation, due in 2029, is the natural home for it.
By the time of that review, an investor should be able to see which providers did what they claimed to do, how much estimation sat behind their scores, and how quickly they corrected course when the facts changed. Today, nobody can.
If standards hold up on their own, the same evidence will show it. That is a test worth running.
Dejan Glavas is professor of finance at ESSCA School of Management.
