En primeur, judged as a transaction.

Every spring, Bordeaux shows the previous year's vintage from barrel, the trade tastes it, scores appear and châteaux release their wine through négociants, often in tranches that usually rise in price as the campaign goes on. Buyers commit and pay now. The wine is bottled and delivered roughly two years later. The same structure appears, less formally, in Burgundy, the Rhône and increasingly elsewhere.

The modern system grew up after the Second World War, when châteaux needed cash long before their wine could be bottled and sold. For buyers the logic was straightforward. Release prices sat below where the wine would eventually trade and for scarce wines the campaign was the only reliable way to secure any at all. Both halves of that logic still hold sometimes. Neither holds automatically.

The arithmetic

You are handing over money two years early. For the trade to make sense, the en primeur price needs to beat the eventual physical price by more than three things combined:

• The cost of your money for two years. Whatever that is worth to you, it is not zero.
• Counterparty risk. You are an unsecured creditor of a merchant until the wine is delivered. Merchants have failed mid-campaign before and customers have lost both the money and the wine.
• Optionality you give up. In two years you will know the bottled wine's actual quality, the market's actual appetite and what else your money could have bought. Committing early forfeits all of that.

Against those, the benefits are: a lower price if the release is priced well, guaranteed allocation where scarcity is real and free choice of format, which is the one benefit nobody disputes.

If the discount to the likely physical price is not obviously large, you are paying a premium for the privilege of paying early.

When it genuinely works

• Genuinely scarce wines. Small production from names with real allocation pressure. If waiting means you will not get any, the calculation changes entirely.
• Strong vintages priced sensibly. Rare and usually obvious when it happens, because the trade says so loudly and stock disappears.
• Large formats. Many magnums and larger bottles are bottled to order. This is the most reliable moment to secure them and they are difficult and expensive to find later.
• Building a vertical. If you are collecting one estate across consecutive vintages, the campaign is the natural rhythm and the completeness has its own value.

When it does not

For most châteaux in most vintages, the honest answer is that you can buy the bottled wine later, in known condition and not necessarily at a higher price, with none of the waiting or the credit exposure. Back vintages of good but unfashionable estates can trade below their own release prices for years.

Be especially wary of a campaign where the case being made is about future value rather than the wine. That is a sales argument, not a tasting note and it tends to appear when the release price is difficult to defend on quality alone.

Reducing the risk if you do buy

• Know who holds your money. Ask whether the merchant pre-pays the négociant and whether your wine is identified to you before delivery. The answer varies by merchant and it matters.
• Prefer merchants who segregate client funds and be sceptical of any offer priced meaningfully below the market.
• Keep the paperwork. Your contract note is the only evidence you own anything for two years. It is also the first document in these bottles' provenance file.
• Do not concentrate. Spreading a campaign across two merchants costs nothing and halves a single point of failure.

Who takes a margin on the way to you

Bordeaux sells through a system called the Place de Bordeaux and understanding it, along with the classifications that rank its châteaux, explains a great deal about pricing.

The château sets a release price and sells to négociants, the merchant houses of the Place de Bordeaux. A courtier or broker sits between them and takes a small commission. The négociant sells on to importers and merchants around the world, who sell to you. Each stage takes a margin and each margin is applied to the one before it. Those margins are one layer of what you pay for in a bottle of fine wine.

Two consequences follow. First, the price you pay is meaningfully above the release price and comparing your quote to the headline release number is comparing different things. Second, when a vintage does not sell, stock accumulates at the négociant level and that overhang can push physical prices below release. It is one reason back vintages sometimes trade under their own campaign price.

The problem with scoring a barrel sample

En primeur scores come from tasting unfinished wine, often soon after the blend has been made and a year or more before bottling. The sample is drawn from barrel, sometimes assembled specifically for the tasting and it is not the wine you will eventually receive.

Good critics say so and publish ranges rather than single numbers for exactly this reason. The market then quotes the top of the range as though it were settled. That gap between what was said and what is repeated is where a lot of poor buying decisions live. Vintage reports written early invite the same mistake.

Three practical rules follow:

• Treat a barrel score as an estimate with error bars and treat a range of 94 to 96 as a claim about uncertainty, not a claim of 96.
• Check the same critic's in-bottle score two years later before you buy the next campaign from the same estate. The pattern of who tends to revise up and who revises down is informative and free.
• Discount any commentary that leads with market performance rather than with the wine. It is telling you what it is.

None of this makes the system dishonest. It makes it a market with information asymmetry, which is a thing anyone who has worked in one should recognise and price accordingly.

The honest summary

En primeur is a good mechanism for scarce wine, large formats and completeness and a mediocre one for everything else. It is not a way to make money and any campaign sold to you on that basis should raise a question rather than an order. If the wine then crosses the world, the real landed cost belongs in the calculation from the start.