In 1855, ahead of an exhibition in Paris, the Bordeaux wine trade was asked to produce a ranking of the region's best estates. They compiled one, based largely on the prices those estates had been achieving over the preceding decades.
It was a snapshot of a market. It was never intended as a permanent hierarchy. And it is still, essentially unchanged, the organising structure of the most commercially important fine wine region in the world.
What it did
The classification divided the châteaux of the Médoc, plus one from Graves, into five tiers, from first growth down to fifth. A separate classification covered the sweet wines of Sauternes.
The methodology was straightforward and quite sensible for its purpose: brokers used the prices each estate had commanded over many years as a proxy for quality. It was, in effect, a market-derived ranking rather than a tasting-derived one.
Critically, the classification attached to estates — to properties and their names — rather than to specific parcels of land. That distinction turns out to matter enormously.
Why it hasn't changed
In its entire history there has been essentially one substantive alteration, when one estate was promoted from second to first growth in the 1970s after decades of campaigning.
Otherwise the ranking stands. Estates that have declined remain classified. Estates that have improved dramatically remain where they were placed in 1855.
The reason is economic. Classification status is embedded in land values, in bank lending, in inheritance valuations, and in the price of every bottle. Demoting an estate would destroy a substantial amount of value overnight, and the owners of classified properties have both the incentive and the influence to prevent it.
Any revision would create losers with more to lose than the winners have to gain, which is a reliable recipe for stasis.
The boundary problem
Here's the flaw that's least appreciated. Because the classification applies to estates rather than to defined plots of land, and because estates have bought and sold vineyard parcels continuously for 170 years, the physical land behind a classification has changed substantially.
A château classified in 1855 on the strength of a specific set of vineyards may now own quite different land. Parcels have been sold, others acquired, boundaries redrawn, estates split and merged.
So the ranking describes the reputation of a name rather than the quality of a site, and the two have drifted apart over nearly two centuries.
What it gets right anyway
Despite everything, the classification remains a reasonable guide, which is genuinely surprising.
Study after study comparing the 1855 hierarchy to current pricing finds substantial correlation. The first growths are still the most expensive. The broad tiers still roughly track quality.
Part of the explanation is that the original methodology captured something real about vineyard sites — the best gravel banks with the best drainage were producing the best wine then and are still doing so.
The other part is self-fulfilling. A classified growth commands higher prices, which funds better viticulture, better winemaking, more selection and more new oak. The classification creates the resources that justify it.
The anomalies
The interesting cases are the ones where it clearly fails.
There are fifth growths that consistently outperform second growths and price accordingly, and second growths that trade well below their rank. Anyone who follows Bordeaux knows which these are, and the market has quietly reclassified them without any official change.
The larger omission is everything the classification never covered. The right bank — Pomerol and Saint-Émilion — wasn't included at all, and it now contains some of the most expensive wine in the world. Pomerol has no classification whatsoever, which hasn't prevented its top estates from achieving extraordinary prices.
That's the strongest evidence that the classification isn't necessary for a market to identify quality. It just happens to be the structure the left bank inherited.
The Saint-Émilion alternative
Worth mentioning because it demonstrates the other approach. Saint-Émilion has a classification that is revised roughly every decade, with estates promoted and demoted.
This should be better and in practice it has been chaotic. Revisions have been litigated, results annulled, and in recent cycles several leading estates withdrew from the process entirely rather than participate.
The lesson seems to be that any classification with real economic consequences will be fought over, and that a static unfair one may generate less conflict than a dynamic contested one. Not an inspiring conclusion.
How to use it
Practically: treat it as a rough guide to price rather than to quality, and be aware that the best value in Bordeaux is generally found precisely where the classification is wrong.
The lower classified growths that punch above their rank, and the unclassified crus bourgeois that outperform their status entirely, are where a bottle costs less than the liquid deserves. Everybody in the trade knows which they are and the information is not hard to find.
Meanwhile the first growths remain excellent and are priced as scarce luxury goods rather than as wine. Which they have been, more or less, since 1855.