How much does it cost to make great wine?

My article in the Financial Times on 15 September is reprinted below, along with a response from Allan Sichel of the eponymous house in Bordeaux, with an interest in several Bordeaux châteaux. An Australian reaction also follows... and another from the wine writer I find more admirably sympathetic, Gerald Asher of Gourmet magazine. And now yet more feedback from Michael Banton of Chateau Rouquette.

Much of my last year or two has been spent hunched over keyboard and map, revising and updating The World Atlas of Wine, the classic wine reference book launched by my friend and colleague Hugh Johnson 30 years ago in a wine world dominated by hock and sherry.

I expected to be daunted by the task: overwhelmed at first, bored in the middle and downright incredulous at the end. I did not expect to be surprised, for in theory I was supposed to be passing on what I had already learnt elsewhere.

But Hugh had introduced a small section entitled 'A note on costs' and it was my job to update it for this fifth edition. Since Bordeaux is by far the most important and best documented fine wine region, and the figures used are based on those of Bordeaux's official agricultural accountants, I moved this analysis from the general introduction to the Bordeaux section and started to examine the latest figures.

I was amazed once I reduced the calculations to per bottle figures by just how little it costs to make even some of the most famous wines of the world – by the fact that it costs next to nothing to make basic AC Bordeaux.

As shown in the table, the average cost of making enough classed growth red bordeaux to fill a standard 75cl bottle is just Ffr28, less than £3 or $5. Nowadays it is common for wine drinkers to be asked to pay ten times that for Bordeaux's more sought-after wines, a year before they even reach the bottling line.

These official figures have been calculated assuming a thoroughly traditional high-density vineyard planted with vines a metre apart, so 10,000 plants a hectare. They also assume 18 months in expensive new 225-litre oak barrels (which can cost more than Ffr3000 apiece but can be re-sold to a lesser property) and an average yield of 45 hectolitres per hectare, the basic yield permitted in a grand appellation such as Pauillac.

Occasionally some natural hazard shrinks the yield to below this maximum permitted level. Many producers cut off excess bunches in summer to keep yields in check. Virtually any decent château owner makes a selection and puts only the best wine under the château's own label, bottling the rest under the label of a second, occasionally third, wine and sometimes selling the rest off in bulk. But the fact remains that most years producers are allowed to harvest an additional 20 per cent of wine above the 45 hl/ha limit, and the average yield of appellation contrôlée wine in the Bordeaux region is always more than 60 hl/ha. The average yield in the unparalleled Pauillac appellation in 2000, for example, was 55 hl/ha.

It is also true that the figures shown in the table are the latest estimates available for what the CGCAG, the Centre de Gestion et de Comptabilité de la Gironde – the official Bordeaux agronomic monitoring agency – would describe as typical properties, rather than the outstandingly good or bad. As this survey is done only every two years, the most recent available for the Atlas were those for 1998.

Furthermore, a typical well run classed growth might cost Ffr28 a bottle to produce, but costs for a first growth or really ambitious second growth could be substantially more than this if teams are sent several extra times through the vineyards to trim surplus leaves and grapes, or not all grapes in a vineyard are picked concurrently. It costs about Ffr4000 per hectare, for instance, to send a team of vine-trimmers through a Bordeaux vineyard.

The so-called 'garage wines' so fashionable on the right bank in and around St Emilion and Pomerol that are made in tiny quantities and with vast amounts of manipulation and human input (each barrel practically has its own personal midwife) may cost even more than a first growth to produce.

But if a vineyard is exceptionally placed and carefully tended, then both chemical and human inputs are low, and so therefore are costs. The Thienponts of Le Pin for example, admittedly some of the least rapacious proprietors of a wine that sells for four and sometimes five digits per dozen, told me that they estimate their per bottle production costs are under Ffr50 a bottle.

These estimates of typical production costs take into account labour, materials and depreciation. They do not include the cost of land or anything that happens once the wine reaches the bottling line. So the difference between that cost price of Ffr28 for a typical classed growth and the latest en primeur release prices of sometimes hundreds of francs a bottle is accounted for by packaging, transport, marketing, the middle men's substantial cut, and profit.

It is worth noting how very much more profitable the top end of the Bordeaux wine pyramid is than the bottom, basic AC Bordeaux end. In the great historical scheme of things, the profits available to producers of fine bordeaux is a relatively recent phenomenon. For much of the 20th century even classed growths were run at a loss. But they sure are making up for that now.

All but the last line of these figures, which is based on my own research, are estimates by CGCAG. 

Allan Sichel replies:

Dear Jancis

I have just read your article [in the FT]. The reader could be misled into thinking that the Bordeaux wine industry is an absolute racket and that the tight legislation is only approximately abided by... which it is not!

You rightly indicate that the cost of land is not included, but to give a fairer view it would be necessary to indicate that the cost of one hectare of planted AOC Bordeaux is about Ffr400,000; that planting costs are about Ffr120,000 per hectare over three years and that no revenue will be generated for five years. At top-growth level a hectare planted with vines sells for 5 million francs. If you readjust your calculations taking these values into account you would reach the conclusion that no grand classed growth would be capable of satisfying a shareholder's expectation for a 15% return.

I would also like to indicate that the 45hl/ha yield is the average yield for wines of classed growth level and not an upper legal limit as your article seems to indicate. The fact that châteaux reduce their yield to 45hl/ha reflects their desire to sacrifice volume in favour of quality. The legal limit for Pauillac in 2000 was about 58hl/ha.

Robinson on Sichel (who is the kindest and most generous supplier of statistics on Bordeaux):

While it is true that an increasing number of Bordeaux properties are owned by companies with shareholders who seek a return, there are still many – including many of the well known classed growths and their equivalents on the other side of the Gironde – which have been in the same hands for years and have no need of amortisation or shareholder dividends.

As for yields, the official regulations for the appellation Pauillac for example cite specifically that the 'rendement de base' is fixed at 45hl/ha and that the 'plafond limite de classement' is just 20 per cent, suggesting that 54hl/ha is a workable maximum yield in the best of years. If the legal limit for Pauillac in 2000 was 58hl/ha, practically all growers must have pushed pretty close to it if they managed an average of 55hl/ha.

But I'm prepared to accept Allan Sichel's suggestion I need to go to Bordeaux to investigate this further. The four Sichel brothers who have succeeded their unusually wise father Peter A Sichel in looking after the family firm seem still to have the odd bottle of Ch Palmer 61 in their cellars... 

And here's further input from Brian Fletcher, chief winemaker at Calatrasi in Sicily, involved among other things with Tunisia's first promising bottled wine exports (stocked in the UK by Majestic Wine Warehouses):

Dear Jancis,

I saw your article regarding the cost of wine in Bordeaux and this is a further insight into relationship between land price and bottle price. In Australia it's a critical success factor to clearly relate the capital/running costs with those of final wine quality/price. So I was prompted to send you this piece from the wires in respect of Codorniu buying vineyards in the Napa. I am not sure if it was with a crop, either way.... 'Spanish cava and premium wine producer, Codorniu has completed the purchase of 74 hectares of vineyards in Foss Valley situated in Napa Valley, California, for a sum of $7.42m'.

This gives us US$40,000/acre. Production of say 4 tons/acre, ie, 260 cases/acre with the average retail price of $60/bottle or average FOB winery of $360.00/case to the trade, neglecting cellar door sales. This gives us average revenues/acre of $ 93,600, with average farming cost of $ 3500 + financing cost $8000 & cost of $10/bottle or $120.00/case.

Quite a profitable business.

I am working here with two other Australian winemakers exploring the potential of Nero d'Avola in Sicily, Primitivo in Puglia and Carignan in Tunisia. However, what strikes me here is the level of EU subsidies granted to growers for just about anything, that in the end as far as I can see has produced nothing of value to the longterm viability and competitiveness of this part of the world.

Gerald Asher adds:

Allan Sichel is right in the main thrust of his argument, and I have no objection to classed growths selling, even at the property, for ten times the raw cost. I don't expect to get a Huntsman suit, after all, for the cost of the fabric and thread. But what irritates me beyond words (and does you too I suspect) is the sheer silliness of cult/garage prices and their effect on the market in general. This phenomenon has little or nothing to do with connoisseurship, the aesthetic of wine, or even sensual pleasure. Bordeaux will be unable to back down from its opening prices for 2000, unfortunately, and we are facing a somber period when even those who can throw money around will hesitate to do so. At least the wines are good and will not be a dead weight on the market as the 1984s were.


And a more recent note from Michael Banton who has owned Chateau Rouquette in the basic Bordeaux appellation for the last four years.

Dear Ms Robinson

I have read your articles and books over the years with great enjoyment. I was particularly interested by your article in the Financial Times of 16 September 2001, on the costs of making Bordeaux, as I bought a vineyard four years ago making Bordeaux Supérieur. It is unusual to see a wine writer looking at the economics of production.

Whilst I entirely agree with much of the article, in particular the high margins of the top wines, I think the small reference you make to basic AC Bordeaux is misleading. You state that it 'costs next to nothing to make', at FF 5 per bottle and that the 'typical ex-cellar price per bottle' is Ffr11. Whilst I think your cost price is about right, this selling price is much too high today. It creates the impression of a very profitable business. This is at odds with the local press which continually refers to the makers of AC Bordeaux as being 'in crisis'. Since this comprises the greatest number of vineyards in the Gironde, with the top wines a minority, it is important for the region as a whole.

Whilst my own experience would confirm the Ffr5 production cost as being about right (our own costs have increased well beyond that since we bought our property and started making improvements). However, I think the selling price of Ffr11 is misleading. It is not clear exactly what this represents. I assume it is a mixture of wine sold to cooperatives, to negotiants en vrac and to wholesalers in bottles.

If one takes the example of sales en vrac, I suspect the reason your selling price is so high is because it represents the exceptional year, 1998. This was when prices briefly went mad for the 1997 vintage. I attach a graph I keep showing monthly en vrac prices for the latest vintage of Bordeaux Supérieur. You will see how it briefly exploded in 1998 but has since returned to the levels it has been at for many years.

The average monthly en vrac price for Bordeaux Supérieur was Ffr10,588 a tonneau in 1998, equivalent to Ffr8.80 a bottle. When you factor in the higher price for bottles, you could get close to your Ffr11 figure. However, bottling and labelling costs about Ffr4 a bottle, which is not apparent from your table – it would increase the cost from Ffr5 to Ffr9 per bottle.

En Vrac prices

The current en vrac price for basic red AC Bordeaux is Ffr6-7500 a tonneau: equivalent to Ffr5-6.25 a bottle! Bordeaux Supérieur is selling for Ffr7500-8500 a tonneau and the white wine is selling for half this. This is a long way from your Ffr11 figure. If costs are Ffr5 a bottle, then the profit per bottle on basic red Bordeaux wine is between Ffr0- 1.25 per bottle. This explains the feeling of crisis in much of the Bordeaux wine trade, outside the grands cru.

Of course, the reason the selling price of basic Bordeaux is so low is because there is a lot of very poor wine around under the appellation name. However, that is another story!

Postscript from Michael Banton on learning I had asked Waitrose's Bordeaux buyer for the ex-cellars prices of AC Bordeaux and AC Medoc.

I understand now the missing factor in the numbers in the table. Waitrose will presumably be paying Ffr11 for a bottle, and not buying the wine in bulk and paying for the bottling. The table in your article did not include bottling, corks, labelling, carton and capsule costs. In our case this totals about Ffr4 a bottle. This would raise your cost from Ffr5 to Ffr9 a bottle and cut the profit from Ffr6 to Ffr2 a bottle.

However, I think Ffr11 is very low, and I wouldn't have thought that you need to adjust that down – certainly, we sell for more than that. There is a small, but growing, number of excellent Bordeaux AOCs who have invested and taken the risks necessary to go up market. (There is some excellent, reasonably priced wine in this category, which rarely seems to appear in the UK.) The problem in the Gironde is for the bulk of vignerons who sell in bulk to negociants or, even less profitably, to the coop (about 25-30% sell to coops).

You need not make your readers feel sad on our behalf: we bought the vineyard four years ago with eyes wide open! It certainly wasn't for financial gain but out of a love for wine and to experience a wonderful way of life. We have no regrets, quite the contrary! Seeing the wine improve is the reward we wanted!

You asked about our harvest. We finished harvesting our Merlot last week and the initial impression is a very positive surprise. It may even be better than our 2000. However, we still have our Cabernet to pick and the acidity is refusing to come down. We're biting our fingernails, checking the grapes twice a day, and praying for the two or three hot days we need! All our neighbours seem to have started harvesting the Cabernet.

See also the report on the 2001 harvest in Bordeaux from Jean-Michel Cazes of Ch Lynch Bages.


Typical production costs in Bordeaux (Ffr)
  AC Bordeaux AC Médoc Classed growth
Number of vines per ha 3,333 6,666 10,000
Harvest costs per ha 2,181 5,192 10,125
Total viticultural costs per ha 35,963 51,533 116,910
Yield (hl) per ha 62 55 45
Total viticultural costs per hl 580 937 2,598
Vinification cost per hl 47 55 195
Barrel ageing (6 months)   316  
Barrel ageing (18 months)     988
Total production costs per hl 627 1,308 3,781
Total production costs per bottle 5 10 28
Typical ex cellars price per bottle 11 22 100