FTSE’s Own FAQ Lists Six Research Languages in English and Seven in Japanese

Ratings methodology is not usually a translation story. This one is.

FTSE Russell publishes a FAQ for the companies it researches — ESG Scores and Indices – FAQ: Information for researched companies. The current edition is v1.5, December 2025. Question 1.8 asks: “Do you only use English information?” For Thai ESG advisory work, ours included, it is the most-quoted line in the entire FTSE corpus, because it is the only place FTSE Russell says anything at all about language.

On 26 August 2026 we pulled that FAQ in both its English and its Japanese edition and read question 1.8 side by side. They do not say the same thing.

We read the same FAQ twice, in two languages

Both files carry the same version number and the same month on the cover: v1.5, December 2025. Same document, same team, same purpose. The only variable is the language of the reader.

The English edition answers question 1.8 like this, quoted in full:

Six languages. A closed list, with a full stop at the end. Now §1.8 of the Japanese edition of the same v1.5:

主な対応言語は、英語、フランス語、スペイン語、ドイツ語、ポルトガル語、日本語、中国語等です。

Rendered into English: “The main supported languages are English, French, Spanish, German, Portuguese, Japanese, Chinese, etc.”

Six in English, seven in Japanese, and an “etc.”

Two differences, and both of them matter.

First, the Japanese edition includes Portuguese. The English edition does not. Second — the larger one — the Japanese sentence ends in 等, the standard Japanese marker for “and so on”. That one character converts a closed list into an open one. The English edition carries no equivalent word, and no equivalent hedge.

The explanation is almost certainly mundane. FTSE Russell’s previous FAQ — the ESG Rating FAQ for Constituents, March 2022 — answered the same question with a seven-language list that did include Portuguese. Somewhere between 2022 and December 2025 the English edition was revised and Portuguese was dropped. The Japanese edition appears to have carried the older list forward unamended. This is a documentation-maintenance gap, not a conspiracy.

But the consequence is real: which version of the rule you believe depends on which language you read it in. That is precisely the failure mode we are retained to prevent in the opposite direction — in a 56-1 One Report, a sustainability report, a green bond framework. Version drift between language editions is not a cosmetic problem. It is a factual one, and here it has happened to the rater.

What this does not mean

We want to be careful here, because this finding gets overstated across the Thai market — and we have overstated it ourselves in the past.

It does not mean FTSE Russell refuses Thai-language disclosure. No FTSE document says that. The binding methodology — the ESG Data Model Methodology v1.2 — is silent on language entirely, as are the FTSE4Good Thailand Index Ground Rules, which are the most Thailand-specific FTSE documents in existence. The language list lives only in a non-binding FAQ, and one edition of that FAQ explicitly leaves it open-ended. The word is “key” languages, not “only”.

It does not mean Thai companies are being marked down. The opposite is on the public record. SET’s published figures for the 2025 assessment round show 222 Thai listed companies averaging 3.6 out of 5.0 on the FTSE ESG model, up from 3.0 the year before, with Governance averaging 4.5. Thai scores are rising, not being suppressed — and any client can find that in one search.

And there is no study behind the strong claim. There is no peer-reviewed research establishing a language penalty in ESG ratings. We have searched for it repeatedly and it does not exist. Anyone citing one is citing something that is not there.

What it does mean: a sourcing risk, not a scoring penalty

So what is the honest version? It rests on how the score is actually assembled.

Put those together and the exposure is evidentiary, not punitive. An indicator is met only if a researcher finds the disclosure and credits it. Evidence sitting only in a language that the research process does not name among those it covers is likelier to go unfound — and your remedy is a two-to-four-week window in which you must hand over a URL and a page number.

That is the defensible claim, and the only one we make: a covered-language edition reduces the risk that disclosure you have already made goes uncredited. It does not “earn” the score, and it does not raise a score by itself. Language is not an indicator. The English edition is simply the version most likely to be read — and that is enough of an argument without inflating it.

Where a hard language gate really does exist

There is one point in this chain where a language rule is not a matter of interpretation at all — and it is not in FTSE’s own documents.

In the same FAQ, FTSE Russell states plainly that “CDP information is material to a company’s ESG Score.” CDP, in turn, sets a hard gate in its Terms of Disclosure, §12.3:

Thai is not on that list. A Thai-language CDP response is not scored — and the resulting gap feeds a rating input that FTSE Russell itself calls material to your ESG Score. Unlike the FAQ language list, this is a published rule with an explicit stated consequence.

So the sharpest language exposure for a SET-listed company is not the One Report at all. It is the CDP response, where the rule is written down, the consequence is automatic, and Thai is expressly outside the set. If you report to CDP, that is the first place to look.

Two clocks, and one that has already stopped

While we were in the primary documents, we checked two mechanics that catch out almost every Thai issuer we speak to.

The first is a fiscal-year trap. LSEG’s Guide to FTSE and Third Party Sustainable Investment Data (August 2026) states that a company’s annual report, sustainability report and other ESG information “needs to be published with the same fiscal year for an ESG score to be updated ahead of an index review.” Thai issuers routinely file a 56-1 One Report for one year alongside a standalone sustainability report still covering the prior year. On FTSE’s rule, that mismatch means the score is not updated at all — the review is forfeited silently, with no failure notice.

The second is that Climate runs on a different clock. The widely-quoted deadline is 30 September, the last date a disclosure can be published and still count for the December index review. That is correct — and for December 2026 it is about five weeks away. But it does not govern the Climate theme. Climate is the one theme of the fourteen not scored by counting indicators: FTSE derives it from the TPI Management Quality score plus emissions relative to peers, and LSEG states MQ scores “are updated annually and are based on reports published by the last business day of March.” Anything you publish in September moves your other thirteen themes. It cannot move Climate.

LSEG also confirms that FTSE Russell indices using MQ data migrate from version 4.0 to version 5.0 of the TPI MQ methodology in September 2026 — this month. Climate scores can therefore shift on methodology alone. If you sit in a Primary Impact subsector, this is the theme with the genuine hurdle, and independently verified Scope 1 and 2 data is the gating item.

What to do before 30 September

The FTSE4Good Thailand thresholds are published and specific: an ESG Score of 2.9 or above to be added, 2.4 or above to stay in, and a Climate Change theme score of 3 for Primary Impact subsectors — cement, chemicals, refining, iron and steel, conventional electricity — or 1 for Secondary Impact, which covers banks, property and most REITs.

We publish this one slightly against our own commercial interest. The strong version of the language argument — that Thai disclosure is ignored, that a theme thin in English scores zero — is not supported by FTSE’s rulebook, and we no longer make it. What the rulebook does support is narrower and more useful: FTSE scores you from what its researchers can locate in the public domain, on a fixed calendar, with a short and unextendable right of reply. Everything you publish is either findable evidence or it is not.

That is a document-control problem before it is a translation problem, which is exactly why the two belong in one workflow. See how we clause-map the two editions → · compare what each rater actually says about language → · or start with the ISO 17100 process behind the English edition →.

Sources, all read at primary source on 26 August 2026: FTSE Russell, ESG Scores and Indices – FAQ v1.5, December 2025, English and Japanese editions, §§1.6, 1.8, 1.9, 1.19; FTSE Russell, ESG Rating FAQ for Constituents, March 2022, §11; LSEG, Guide to FTSE and Third Party Sustainable Investment Data used in FTSE Indices, August 2026, §2.8.3; FTSE Russell, ESG Data Model Methodology v1.2, June 2026; FTSE4Good Thailand Index Ground Rules v1.0, June 2026, §6.3; CDP, Terms of Disclosure, §12.3. Thai score distribution as published by SET; FTSE Russell has not published a Thailand-specific comparison.