One of the main statistical canards about wine is that smaller yields equal better quality.

This is so widely believed in Europe that wine seminars become predictable. If a winemaker is on a podium talking about a wine, the first question is never about philosophy or intention, but “What was the yield?”

W. Blake Gray

The theory is that when a grapevine has to concentrate its energy into fewer grapes, those individual grapes will be more concentrated and packed with flavor. So wines of lower yield should taste better.

This idea leads winemakers at high-end wineries to slice off bunches of healthy, immature grapes and leave them on the ground, cutting their potential crop in half.

Farmers of other crops would never do that. But fine wine sells for so much money that if a winery can charge more than double by making half the wine, it makes sense economically. In Napa Valley, winery owners brag about their low yields. “You got 1.5 tons an acre? Well, we got only 1.3.”

But what if they’re wrong? What if lower yield doesn’t equal better quality? What if all those grapes left on the ground would have been worth just as much as the ones that ended up in precious hand-etched bottles?

There’s a challenge to testing any relationship between quality and quantity in wine: quantity is easy to measure, but quality is not. To use critics’ ratings, you have to agree with the personal taste of the critic, and I think we all know how Palate Press readers feel about that. But if we don’t use scores, what statistical measure can we use?

Channeling my inner Adam Smith, I hit upon wine prices in the secondary market. The secondary market includes older wines that are either being sold by a retail store that has been holding onto them, or are being resold by their original purchaser.

It is imperfect, because Robert Parker’s ratings are important in secondary market prices. But it is at the very least a measure of what the market is willing to pay for a wine.

I created an index of the ten California wines most searched on Wine-Searcher. Conveniently, nine of these wines are from Napa Valley, and the tenth is a special case I will address below. 10 most sought wines

10 most sought winesWhile quality will always be on the palate of the drinker, averaging the prices of the most-sought wines by people who seek wines should make the market value of these wines a meaningful measure of their perceived quality.

I used the overall Napa Cabernet harvest, amount harvested divided by acres planted, as a figure for yield. Of the 9 wines made in Napa, four are single-vineyard wines, but 5 are made from vineyards scattered around the valley, so the county-wide yield is applicable. It also should be relevant to the four single-vineyard wines because general weather conditions are very similar throughout the valley.

I plotted secondary prices versus yield over the ten years from 1997 to 2006; wines that the market has had a chance to pronounce judgment on.

What I found is that in California, there is actually a  positive relationship (R = 0.22) between yield and secondary market price: wines from higher-yield years are likely to cost more. The relationship is weak, indicated by R2 of 0.049. But it’s also the opposite of what conventional wisdom would expect.

combinedIt’s worth pointing out that in high yield years, by definition, there is more wine to sell, and the law of supply and demand should drive prices down. But that isn’t happening.

Look at the lowest-yield years on the chart. The 2004s, easily the scarcest, have the same resale value as the much easier-to-find 2003s, and cost almost the same as the abundant 2005s. It’s not what you’d expect to find either by the rule of scarcity, or if lower yield equals quality.

Let’s look closely at that 10th wine, Ridge Monte Bello.

It’s an excellent wine to investigate in depth because in an opaque industry, Ridge is unusually transparent. Ridge lists not only all the grapes that go into its wine on its labels, but also any added ingredients like calcium carbonate, water or SO2.

Ridge Monte Bello is the 7th most searched American wine on Wine Searcher. It’s consistently popular and well-regarded critically, and its yield jumps all over the place: more than 5.5 times as much was made in 1997 as in 2002. Supply and demand might make you expect the ’02 to be rarer and more expensive than the other wines, but it’s in the middle of the pack price-wise, whereas the copious amounts of 1997 Ridge Monte Bello available don’t prevent it from being the second-highest priced.

Once again, there is a weak correlation: an R of 0.1549, less strong than for the Napa wines. But once again, that correlation is positive: the more wine Ridge made from its Monte Bello vineyard in a given year, the more that wine is likely to be valued by the secondary market.

Now just think of all those valuable grapes left on the ground over the years.

About The Author

W. Blake Gray
Staff Writer

Wine writer W. Blake Gray is Chairman of the Electoral College of the Vintners Hall of Fame. Previously wine writer/editor for the San Francisco Chronicle, he has contributed articles on wine and sake to The Los Angeles Times, Food & Wine, Wine & Spirits, Wine Review Online, and a variety of other publications. He travels frequently to wine regions and enjoys coming home to San Francisco.

Related Posts

12 Responses

  1. Paul Franson

    Research at UC Davis seems to indicate that quality vs. yield on an individual grapevine is a bell-shaped curve. Reducing below a certain level doesn’t increase quality, i.e., 3.5 tons per acre for Cabernet at a certain site in Napa may be best. At 2.5 tons, you just get less wine. At 6, the quality may suffer. The optimum appears to be higher in rich valley soils, lower in the hillsides and mountains. The specifics depend on the variety, cultivar and rootstock, location and weather, of course. But perhaps someone there could say more.

    Growers of Muscat of Alexandria in the southern Central Valley may get 30 tons per acre with “suitable” quality. That is, it’s no better at 20…

  2. Kyle Schlachter

    Blake, while I applaud you for trying to do something interesting, your methods and conclusions are tenuous, at best. First, you use try to find a connection between two disparate populations. Perhaps contacting the wineries for yield data would have made sense. Second, your data indicate (r=0.22, N=10, p<0.1) absolutely no correlation between price and yield. Your correlation coeffcient would have to be greater than 0.549 to be even somewhat significantly correlated (α=0.1). Next time, Brush up on statistics before you use them.

  3. Tyler R Thomas

    While Kyle makes a good point, your conclusions are not off. That is, the lack of correlation would be expected from anyone who has knowledge of vine physiology. I would argue for at least two decades now we have known that there is no linear relationship between yield and quality even though high quality does often come from vines that produce less fruit. But that yield (especially tons per acre) can forecast quality is known to be dubious. This perspective was built within the scientific community with peer reviewed study. However in the wine business the scientific community does not control the narrative. What matters most is what people believe and even many winemakers still believe in the relationship because as Paul noted, there are limits on the low and high yield end that do begin to cause changes (or lack thereof) to quality.

    Broadly speaking I think it is safe to say that many believe lower yields have better quality, and that smaller berries have better quality. What many do not realize, is that if you increase yields on a single vine often the berry size actually decreases. Conversely, one way to get larger berries is to reduce yields significantly early in the season. But if smaller berries come from higher yields, how can one factor improve quality and the other decrease it? All that and we haven’t even talked about assumptions of between row spacing which can drastically change tons per acre without even altering the load per vine. If your grams per vine are the same and you go from 10 foot to 5 foot spacing, you’ve doubled your tons per acre without altering the physiology of the vine. And it is the physiology of the vine that we ultimately should concern ourselves with.

    Of course low yielding vines do often produce better wine. However the question should be why are they low yielding? Is it the low yields that produce the better wine, or are low yields a concomitant end result of whatever is impacting the vine’s physiology such that quality improves? If I cut off half my fruit from my valley floor vineyard to obtain the 3.75 tons per acre that my neighbor’s hillside vineyard achieves (on the same row spacing) but I still irrigate too much and my neighbor doesn’t irrigate at all; I may have the same yield but his wine will be much better.

    What journey did the vine take to obtain the yield it provides? That is the most important question.

    • Darius Liddell

      this is the most important comment on this article.

  4. candid wines

    We asked Kathleen Inman about lower yields and Pinot Noir for Ask a Winemaker and she felt like the question was a softball…

    In my importing / distribution business, both Sean O’Callaghan at Riecine and Francesco di Fillipis at Cosimo Maria Masini (both in Tuscany) talk about the importance of cropping Sangiovese higher in order to reduce the concentration of the fruit on the vine, avoid runaway alcohol and preserve the freshness of the variety.

  5. SiduriWines

    A few things I’d wonder about:

    1) The grape crush report (which you used to get tons crushed) actually measures grapes purchased by wineries and made into wine. Yields in those cases could be higher I assume (growers doing what they can to get higher yields etc) than it might be for estate wineries. So which of these wineries are Estate in their grape sourcing?
    2) Planting density….Tyler alludes to this. Places like Opus are pretty densely planted and thus yields on a TPA basis might be higher but not on a lbs per vine basis.
    3) I wonder about the drop off/consistency of price of all of the wines from 2003 forward. That seems very odd given the variability of the preceding period. I wonder where the secondary market prices came from? Has the market not had enough time to make a judgement on the wines from 2003 onward?
    4) Per Candid Wines comments below — if you look at the wines from 1997 through 2002 one thing you can see, I think (I am not a statistician), is a relationship between average brix and pricing. The 3 highest average brix years were 1997 (24.5), 2001 (24.8) and 2002 (25.3). Those are also the wines that are priced the highest…..

    So, Blake, let’s see about changing the title to “Does high brix mean higher quality?” 🙂

    Adam Lee
    Siduri Wines

  6. Steve Pogue

    Here’s a link to a recent article looking at sugar, pH and anthocyanin levels for Cab by price/ton. If you look at the yield ranges, price/ton is a decent proxy for overall vineyard yield. The study found no correlation between those 3 juice parameters and the price of grapes. The lead author, Mark Matthews at Davis has done a lot of research to dispel the yield/quality dogma. You may also want to talk to Mike Anderson, who runs the UC Experimental Vineyard in Oakville, which sells most of its fruit to wineries with high bottle prices. Mike is a grower from a family of growers and I believe he’d echo what other posters have said – that fruit quality is optimal over a range of yield. That range depends on the varietal, site, disease pressure and other factors, not an arbitrary number that looks good on the tasting notes.

    • Tyler R Thomas

      Thanks Steve! I came out of the Matthew’s lab and my comments were certainly informed by his work.