Corrections

This register was audited before a single page of it was generated. Sixty-one checkable claims went to their sources. Thirty-eight survived. Twenty-three did not, and a further eleven were stating a figure, a date, an ownership or a rank with nothing on the page to say where it came from.

No page may state a figure, a date, an ownership or a rank unless the audit marks that claim verified, or the page itself names the publisher, the table and the year on its face. A number with no source does not get softened. It gets cut.

Nothing below was corrected by inventing a replacement. Where nothing verifiable existed, the figure was removed and the sentence around it was rewritten to stand without it. Five figures were cut outright and nothing was put in their place. That is the point of this page.

How to read it

Everything struck through in this shade is what the register used to say. It appears nowhere else on this site: the build greps all one hundred and sixteen pages for every retired string and refuses to publish on a hit, and this page is the single exemption, because quoting what the register used to say is its whole job. Every entry carries the date it was changed and the source that changed it.

1. Cut, and not replaced

1.1 The luxury table that does not exist — 23 September 2026

There is no standalone Luxury & Premium 50 for 2026. None was issued that year. The register's own méthode page said as much, and four claims elsewhere in it cited that table anyway. Two of them were display copy, on the homepage and in a category essay.

Source: Brand Finance, Luxury & Premium 50 2025 — brandfinance.com and across-magazine.com.

The build now refuses to publish that publisher's name within forty characters of that year, unless the word immediately after it is Apparel or Cosmetics. Those two tables do exist for 2026, and the register uses them correctly.

1.2 The six-way country split of luxury brand value — 23 September 2026

The dossier carried France ~48, Italy ~18, Germany ~12, Switzerland ~9, US ~7, UK ~4. It sums to 98% and it is published nowhere. Cut. What Brand Finance publishes is one sentence, and that is what the site now says: France contributes 49% of total brand value, nearly triple Italy's share (Luxury & Premium 50 2025). The homepage stat card that read 47.7% now reads 49% and names the table. The country notes for Italy and Switzerland lost their shares of the same unpublished split — ~18% of luxury brand value and ~9% of luxury brand value — and France's note lost ~48% of Brand Finance.

1.3 "Automobiles 21% of the table" — 23 September 2026

Apparel's share is published: 32 brands and 70% of total value. The matching figure for automobiles 21% is not published anywhere. It is cut; the apparel figure stays and is attributed.

1.4 The Birkin resale multiple — 23 September 2026

Resale of a Birkin 25 routinely clears 2× list rests on no published index. Resale spread varies by leather and hardware and nobody indexes it. Cut. In its place, a figure that is published: a Birkin 25 lists at $13,500 in the United States after the January 2026 increase, up from $12,700. Source: PurseBop.

1.5 Loewe, the Only Spanish name in the global luxury 50 — 23 September 2026

A placement claim that names no table and no year, and no published luxury fifty was found that supports it. Cut, with no substitute. Loewe's page carries no citation block.

1.6 Three of the four labs.llc stat cards — 23 September 2026

The rooms page claimed 13 rooms / 20 doors / 54 live-data builds / 167 assets. Only one of those four is countable from this site's own data. Twenty doors stays, and gains a fact this build actually measured: all twenty answered HTTP 200, with zero redirects, on 23 September 2026. The other three were labs.llc internals this build did not count, and they are gone.

2. Ownership — five parents were twelve to twenty-four months out of date

Ownership is the claim a register gets asked about, and the one that ages fastest. All five were wrong or stale on the day the audit ran. Each now carries an as-of date on the page.

Three of these were repeated in prose, which is worse than a wrong field, because a field can be swapped and a sentence has to be rewritten:

• Off-White. The essay's closing line was LVMH’s job is to keep the language from becoming a costume — a job LVMH stopped having two years ago. The sentence is rebuilt around the sale: the language is a licensing house's to keep honest now, not a luxury group's.
• Fabergé. The essay tied the house to coloured-stone supply — Gemfields ownership ties the house to the mine. That tie was cut when Gemfields sold, and the sentence says so.
• Creed. now under Kering Beauté became now under L'Oréal, which completed its acquisition of Kering Beauté on 31 March 2026.

3. The atlas — the ordering rule the table was breaking

The hundred-country table said it was Ranked first by the size of the economy. The top twenty matched the IMF's April 2026 World Economic Outlook to the decimal. Below that the ordering fell apart: twenty-two places where a higher rank held a smaller economy, and Taiwan — a roughly one-trillion-dollar economy — sitting at rank 56, below Pakistan at $375B.

The méthode page had an escape hatch for this: Some ranks below ~40 are composed so origin states are not erased. That explains promoting a small craft economy. It does not explain demoting Taiwan thirty-five places.

Decided 23 September 2026: fix the data, not the promise. Every one of the hundred figures was re-read from the source, not just the twenty the audit had already checked:

International Monetary Fund, World Economic Outlook database, April 2026 — GDP, current prices (NGDPD), US$ billions, projection year 2026. Read 23 September 2026 from the IMF DataMapper, whose own metadata names that edition and a last-modified date of 8 April 2026.

Ninety-six of the hundred have a 2026 projection in that edition, and those ninety-six are now ranked strictly by it. Taiwan moves from 56 to 22. The escape-hatch sentence is deleted, and the editorial goal it was reaching for — not erasing the origin states — is served honestly instead: the fifteen countries that host a maison in this register are marked on the table and link to their maisons, wherever their economy puts them.

Four countries have no 2026 projection in the April 2026 edition, so they cannot be ranked against the other ninety-six without inventing a number. They are filed after the ranked table, alphabetically, under a heading that says exactly that:

The table's header claims the ordering it actually has, and no more. The build fails if the ranked block is ever not monotonic in GDP.

4. The bug that printed a false number on every page — 23 September 2026

Rank 68, Mercedes-Maybach, carried a founding year written as 1909 / 2002 in the data file. Unquoted, that is not a date range. It is a division, and JavaScript evaluated it: the house's founding year rendered as 0.9535464535464535 on every card and in every modal the site had.

It now reads 2002, with a note on its face: Maybach 1909; Mercedes-Maybach 2002. Founding years are typed as integers between 1400 and 2026 for all one hundred maisons, and the build fails on any value that is not one. This was fixed before indexing, not after.

5. Corrections of fact

All dated 23 September 2026.

The Limited100 placements — eight of them, all mislabelled

Eight maisons carried a placement in the Limited100 Luxury Brand Index, written as though the index ranked by sector: Limited100 watches No.2, Limited100 auto No.3, Limited100 jewellery No.8, Limited100 rarity index top 15, Limited100 watches No.5, Limited100 auto top 6, Limited100 auto top 11, Limited100 spirits top 12. The index is real and published, and it ranks 1 to 100 overall, not by sector. Patek Philippe is not the second watch; it is the second brand. All eight are restated as No. N of 100 and attributed to Limited100, Luxury Brand Index (L100i), 2025 Q4: Patek Philippe 2, Bugatti 3, Richard Mille 5, Rolls-Royce 6, Van Cleef & Arpels 8, Pagani 11, Louis XIII 12, Balenciaga 14. Source: limited100.co.uk

6. Seven compound origins, split — 23 September 2026

Seven maisons had two countries in a single field. A field used to file a maison by origin cannot hold two of them, and splitting the strings produced the junk token "Japan roots" and a "Hong Kong" that the country table did not contain. Each maison now has one origin and, where the second country was carrying real information, a note that keeps it as prose:

• Balenciaga → Spain. Founded in San Sebastián; the house has been Paris-based since 1937.
• Off-White → Italy. Founded in Milan by the American designer Virgil Abloh.
• De Beers → South Africa. Founded in Kimberley; headquartered in London.
• Fabergé → Russia. Russian origin; the modern brand is owned and operated outside Russia.
• Breguet → France. Founded on the Quai de l'Horloge in Paris by the Swiss-born Abraham-Louis Breguet; the manufacture is in Switzerland today.
• Panerai → Italy. Florentine identity, Neuchâtel manufacture.
• Aman → Thailand, as above.

Three maisons filed under "Hong Kong" — Mandarin Oriental, Rosewood and Peninsula — now read Hong Kong SAR, which is what the country table calls it and what the IMF calls it. Every maison's country is now a country that exists in the table, and the build fails if that ever stops being true. India's note lost Gemfields/Fabergé adjacency at the same time, because that tie ended in 2025.

7. Attribution added where there was none

Eleven claims were not wrong. They were unattributed, which under this site's rule is the same problem with a better disguise. Each now names its publisher, its table and its year on the page itself: Hermès and Fendi (Kantar BrandZ), Louis Vuitton (Interbrand Best Global Brands 2025), Chanel and Dior and Tiffany (Brand Finance 2026 Apparel 50), Gucci and Saint Laurent (Kering FY2025), Prada and Miu Miu and Versace (Prada Group FY2025), Rolex (Morgan Stanley / LuxeConsult 2025), Cartier (Richemont, year to 31 March 2026), and Guerlain — placed No.10 in the Luxury & Premium 50 for 2025 by Brand Finance.

Three claims are the house's own account and are now labelled as such, rather than as fact:

• Creed, founded 1760. Kept, and attributed on its face: 1760, by the house's own account; not independently documented. Historians have not substantiated the London founding date. The house says it, the page says the house says it, and the reader can decide.
• Vacheron Constantin, the oldest watch manufacture in continuous operation, and Boucheron, the first jeweller on the Place Vendôme — both by the houses' own accounts, both labelled.

8. On the watch list

De Beers. The ownership field — Anglo American 85%, Government of Botswana 15% — is correct as of 23 September 2026, and it carries that date. A sale to the Global Diamond Consortium has been announced and is expected to close in Q4 2026. When it does, this page gets a new entry and the field gets a new date. It is listed here now so that the day it goes stale is not the day somebody notices. Source: CNBC Africa

Entrupy. A 2026 State of the Fake exists. This build cites the 2025 edition, because that is the one it read.

9. What was not checked

The audit checked sixty-one claims. It did not check the signature line for any of the hundred maisons, the hundred country readings, the qualitative analysis prose for ranks 26 to 100, the labs.llc estate counts, or the founding years and parent companies one by one. The full list, in the register's own words, is published on the méthode page.

10. How to check this page

The build keeps a machine-readable ledger of every row above — what was said, what is said now, who published it, which table, which year, and the date each was checked — and a list of the retired strings as literals. It greps every page for every one of them and refuses to publish on a hit. This page is the only page allowed to contain them, because quoting what the register used to say is its whole job.

Corrections, disputes and better sources: labs@labs.llc.