Monday, May 29, 2017

How many wine-quality scales are there?

There are a number of ratings systems for describing wine quality, which use 100 points, 20 points, 5 stars, 3 glasses, etc. Unfortunately, there is usually no "gold standard" for these systems, and so no two wine commentators use these systems in quite the same way.

That is, when critics differ in their wine scores for a particular wine, it can be for one of two reasons: (i) their opinions on the wine's quality differ, or (ii) they are expressing their opinion using different numbers. That is, when the critics produce the same score, they may or may not be assessing the wine as having the same quality, and similarly when they produce different scores. Each critic has their own personal version of the "100-point scale" or the "20-point scale".


This situation is similar to people speaking different languages. Simply looking at a word does not necessarily tell you what language is being used, because the same combination of letters can occur in different languages, with or without the same meaning. For example, the word "December" appears in both Swedish and English, and in this case it has the same meaning in both languages. However, the word "sex" also appears in both languages, but in Swedish it usually refers to the number 6, which is not necessarily related to any of the word's possible meanings in English.

So, if the Wine Spectator gives a wine 90 points, does that mean the same thing as when the Wine Advocate gives that same wine 90 points? Probably not. Just for variety, instead of using the 100-point scale to illustrate this topic, I will use the 20-point scale for wine quality — this emphasizes the need to translate the ratings systems to a common one.

20-point ratings systems

Many American wine drinkers are familiar with the 20-point scale developed in the 1950s by Maynard Amerine and his colleagues at the University of California, Davis, intended as a teaching tool for identifying faulty wines. This was, indeed, an attempt to produce a "gold standard" wine rating system. Each organoleptic characteristic of the wine is assigned a number of points based on its perceived quality, and these points are summed to produce the final score. In both theory and practice, everyone who uses the UCDavis scale should be "speaking the same language"; and therefore any differences in wine scores should represent differences in wine quality, not differences in language.

Sadly, not everyone has agreed with or used the UCDavis scale, especially as a general tool for wine tastings; this topic is discussed in detail in recommended books such as those by Clive S. Michelsen (Tasting and Grading Wine. 2005) and Andrew Sharp (Winetaster's Secrets. 2005). So, there are innumerable 20-point scales in use around the world, and they all seem to represent different languages. To illustrate the range of scales in use, we can compare the scores given to the same wines by different critics.

In order to standardize the scales for direct comparison, we need to translate the different languages into a common language. Jean-Marie Cardebat and Emmanuel Paroissien (American Association of Wine Economists Working Paper No. 180. 2015) have suggested doing this by converting the different scales to a single 100-point scale. The one they chose was the scale used by the Wine Advocate (which is not necessarily the same as that used by the Wine Spectator, or the Wine Enthusiast, etc), and I will do the same here. Furthermore, I will compare the quality scales based on their scores for the five First Growth red wines of the Left Bank of Bordeaux (as described in the post How large is between-critic variation in quality scores?).

The scales for five different commentators are shown in the first graph. The original scores are shown on the horizontal axis, while the standardized score is shown vertically. The vertical axis represents the score that the Wine Advocate would give a wine of the same quality. If the critics were all speaking the same language to express their opinions about wine quality, then the lines would be sitting on top of each other; and the further apart they are, the more different are the languages.

Five different 20-point wine-quality ratings systems

Also shown is the difference in meaning for a wine that gets a score of 18 from each of the critics. If we see a wine score of 18, then La Revue du Vin de France, Jean-Marc Quarin and Bettane et Desseauve mean a somewhat better wine than does Jancis Robinson. On the other hand, Vinum Weinmagazin is indicating a somewhat worse wine. They are, indeed, all speaking different languages; and we readers need to translate between these languages in order to get their meaning.

As another example, at the end of June 2012 Decanter magazine changed from using a 20-point ratings scale to a 100-point scale (see New Decanter panel tasting system). In order to do this, they had to convert their old scores to the new scores. They used a conversion that is precisely halfway between the scoring systems of Jancis Robinson and Bettane & Desseauve, as shown in the next graph (see How to convert Decanter wine scores and ratings to and from the 100 point scale). So, this is yet another different 20-point language.

Seven different 20-point wine-quality ratings systems

So far, I have assumed that there is a linear relationship between the scores from the different critics (ie. the graph lines are straight). However, in an earlier post (Two centuries of Bordeaux vintages) I suggested that the relationship between the Bordeaux scores from Tastet & Lawton and from Jeff Leve (the Wine Cellar Insider) is curved, instead. Indeed, The World of Fine Wine magazine explicitly indicates that their 20-point scoring system is non-linear, as shown in the second graph above. This makes for a very complex language translation, indeed.

As we shall see in the next post (How many 100-point wine-quality scales are there?), translating between 20- and 100-point scales is not straightforward, either.

Conclusions

The short answer to the question posed in the title is: pretty much one for each commentator. Fortunately, there are not quite as many wine-quality rating systems as there are languages. Nevertheless, the idea of translating among them is just as necessary in both cases, if we are to get any meaning.

Does all of this matter in practice? Quite definitely. Indeed, every time a wine retailer plies us with a combination of critics' scores, we have to translate those scores into a common language, in order to work out whether the critics are agreeing with each other or not. Since most of us are not doing this, we may well be fooling ourselves into seeing a false sense of agreement among those critics. The world of fine wine is more complex than most people realize, or would like.

Furthermore, this issue is at the heart of the objections that mathematicians have to simply averaging wine scores across different critics. If the critics are all using different ratings scales, then the average score has no mathematical meaning. That is, if the critics are speaking different languages, then what would the "average" of those languages mean? It would be gibberish, unintelligible to anyone, even if the combination of letters looks like it might be a real word. A classic example of this is the Judgment of Paris, from 1976, in which the "official" summed scores are meaningless, because the tasters were all using different versions of the 20-point scale (see A Mathematical Analysis of The Judgment of Paris). Note also, that the scores using the UCDavis scale are much higher than are the scores for the Judgment (see Was the Judgment of Paris repeatable?).

Monday, May 22, 2017

Lazy journalism

This week marks the first anniversary of this blog. This is an important milestone for most blogs; and I have averaged more than one substantial post per week during that time, as I approach 60 posts. By way of celebration, this post is a bit different to most of the others.

This blog usually deals with wine data in the form of numbers, but there are other forms of data that could be used instead. One of these is industry information, as presented by the media. Sometimes, this is more opinion than properly checked information.

Consider this example from The Fabulous Ladies' Wine Society:
Cumulus Wines is a true child of the 80’s. We reckon the owners must have been listening to UB40's Red, Red Wine on repeat on their walkman when they planted out over 500 hectares of vineyard in the barely known Orange region nearly 30 years ago and built a 10,000 tonne winery with storage capacity for 8 million litres of wine. But obviously they were on to something as it has definitely paid off!
Almost everything written there is nonsense, as also is much of what is said about the same winery at Just Wines. So, let's look at the true saga.

Cumulus Estate Wines is itself very coy about the company's history, with its description giving the impression of one continuous flow of time. However, this is far from the truth, as indicated by media reports at the time of the various events, such as those from the Newcastle Herald, The Age, Chris Shanahan (of the Canberra Times), the Pierpont column (of the Australian Financial Review), the Wine Spectator and Wine Genius. The story is long and convoluted, so here goes.


The company's main vineyard area is south of a small town called Molong, which lies just inside the Orange viticultural area of eastern Australia (the vineyard actually straddles the region's border). The vineyard, called Little Boomey, was established by Peter Poolman in 1995, increasing in size to 508 hectares over the next three years. Capital for the development was raised from hundreds of small investors, with the intention that ownership would revert to Poolman’s company after the investors had leased the vines for 15 years. What was then called Southcorp Wines (Australia's biggest wine company) bought and vinified the majority of Little Boomey’s grape harvests.

The Central Highlands Wine Grape Project, as it was officially called, was actually a tax-driven investment scheme with several vineyard areas. It was merged into a new investment company called Cabonne Limited in 1998, which was publicly listed on the Australian Stock Exchange. In 2001, Cabonne took over Reynolds Wine Company (owned by Jon and Jane Reynolds), and changed its name to Reynolds Wines Limited in 2002. It set up its wine making at Cudal, south of Molong, where a high-tech 10,000-tonne capacity winery had been built.

Reynolds Wines soon went bankrupt, slipping into voluntary administration in August 2003. The problem seems to have been what is euphemistically called an "awkward corporate structure", rather than problems with either the winery or the wine business. Notably, the company owed AU$18 million in taxes. The Australian Tax Office was not convinced that the original vineyard schemes were truly tax-deductible — this was not an agricultural development but a financial investment scheme (a decision that they also applied to other vineyard small-investor schemes).

At the time, the subscribers to the original tax-minimization schemes apparently still owned, as license holders, the grapevines on Reynolds' three properties (reverting to Reynolds between 2012 and 2018), and also had rights to the wine made from those vines. The wine was concurrently being sold through a joint venture with the Trinchero group, from the Napa Valley in California (currently the fourth biggest winery in the USA, by case sales). As a result, Trinchero Family Estates acquired the Reynolds and Little Boomey brand names early in 2004, but had no interest in buying either the winery's production facility or its 900 hectares of vineyards.

The bankruptcy receiver (appointed by the ANZ Bank) sold the Cudal winery, the adjacent 508 hectares of vineyard and other assets to Cumulus Wines Proprietary Limited for AU$30 million — much less than the AU$130 million that Cabonne is reported to have invested in developing the property. Cumulus agreed to underwrite the bank loan only, which means that none of the investors got their money back, neither the original grape leasers nor those later investing via stock-exchange shares (ie. the bank came out of this okay but no-one else did!).

The Cumulus Wine company had been set up in 2004 by an underwriter and insurer called Assetinsure Proprietary Limited (50% owned by investment bank Babcock & Brown), based in Sydney. Philip Shaw (former Southcorp head of production) was appointed as the winemaker to develop the new wine company, focusing on cool-climate grapes from Orange and elsewhere in the Central Ranges viticultural area. The Little Boomey vineyard was re-named Rolling. In 2005, Keith Lambert (another former Southcorp chief executive) acquired a 51% stake in the company.

A worldwide distribution network was established. However, this proved to be overly ambitious, in spite of grants from the Export Market Development Grants Scheme, from the Australian government. So, in 2007 the Berardo wine family, of Portugal, bought the 51% share-holding. The Berardo Group has extensive wine investments in Portugal, via the Bacalhôa Vinhos de Portugal group, a 33% stake in Sogrape (Portugal’s largest wine company), 25% of Henriques & Henriques Lda (of Madeira), and joint ownership of Quinta do Carmo (with Eric de Rothschild, of Château Lafite), as well as owning 50% of Colio Estate Wines (one of Canada’s major wine producers).

This partnership between Assettinsure and the Berardo Group lasted for some time; and in 2013 they launched a new wine sales and distribution company, Epoch Wine Group. [Don't worry, you are now well over half-way through the saga.]


However, in 2015 Cumulus Wines was involved in a scrip-for-scrip merger (ie. shares were exchanged instead of cash) with the trading company Wine Insights Proprietary Limited. This company owns Beelgara Estate, from the Riverina viticultural area (south-west of Orange), as well as making wine from the viticultural areas of Margaret River (Moss Brothers label), Coonawarra (Riddoch Run), Mudgee (Frog Rock), Adelaide Hills (Em’s Table), Clare Valley, McLaren Vale and Yarra Valley, among othes. Beelgara Estate was formed in 2001, when a group of shareholders bought the 70-year old Rossetto Wines company, including its winery at Beelbangera, just outside Griffith. This company had then merged with Australian Wine Supply in 2004, and the Wine Insights company was created in 2012, following further acquisitions and partnerships (including contract wine-making and bottling, and bulk wine supply).

The Cumulus merger is reported to have created a joint venture producing, per year, more than 400,000 cases of wine and with a gross revenue of AU$20 million. Winetitles Media now ranks Wine Insights as the 15th largest Australian wine company by revenue (and sales of branded wine) and 20th by wine-grape intake.

However, the venture also put the Rossetto winery, at Beelbangera, up for sale, because the merged group chose to centralize its wine production at the Cumulus winery, at Cudal. This seems to mean that the Riverina grapes are now going to be transported 350 km to be processed (at Cudal), rather than being processed locally (at Beelbangera). Much worse, the Margaret River grapes would be transported 4,000 km for processing, the Coonawarra and Adelaide Hills grapes would be trucked 1,000 km, etc. Environmentally friendly this would not be (with a large carbon footprint), although the accountants must love it.

That's it, for the moment. Nothing stays the same for long in the world of Australia's large wine companies. But the next time you read a media report about some wonderful winery, you should wonder what is the reality behind it.

Addendum 2018

The company has recently reverted to its original name: Cumulus Wines undergoes re-brand.

Monday, May 15, 2017

Opus One, and the argument for varietal diversity

The red wines from Bordeaux contain one or more of several grape varieties: Cabernet sauvignon, Cabernet franc, Merlot, Malbec and Petit verdot. (They used also to contain Carménère, but that grape is now rare in Bordeaux.) When Robert Mondavi and Philippe de Rothschild decided to make a Bordeaux-style wine from Napa-grown grapes, they naturally used these same varieties.

This wine has been known as Opus One, with its first vintage in 1979. It was the first ultra-premium wine from the USA, the California equivalent of a Bordeaux first growth, intended as a benchmark for the wines produced from cabernet grapes in the Napa Valley. It has struggled to maintain that reputation, as it has been persistently criticized for inconsistency from vintage to vintage. Certainly, other wines have surpassed it in price and/or reputation (e.g. Ridge Monte Bello has a similar Bordeaux-style aim), although they all sell considerably fewer than the 25,000 annual cases of Opus One.

This inconsistency bears looking into. I contend that it has at least something to do with the variation in grape varieties.


The wine started out as a blend of mainly cabernet sauvignon, along with some cabernet franc and merlot. Then, malbec was added to the blend in 1994, and petit verdot was added from 1997 onwards. The proportion of these grape varieties in the wine has varied from year to year, as determined by the winemakers. The winemakers were Tim Mondavi and Lucien Sionneau from 1979–1984, and Tim Mondavi and Patrick Léon from 1985–2000, with Genevieve Janssens assisting from 1991–1997. Michael Silacci has been the chief winemaker since 2001, early on with either Tim Mondavi or Philippe Dhalluin, but alone since 2004.

In this blog post I wish to look at the variation through time in the diversity of the grape varieties within the wine. A number of mathematical measurements of diversity have been developed in science, for making precisely this sort of comparison. The idea is to reduce the proportions of the various grape varieties down to a single number (for each vintage) that quantifies their diversity, from a single grape variety at one mathematical extreme to equal amounts of each grape variety at the other extreme.

The one I will use here is called the Shannon Diversity Index (see Wikipedia). This Index will be a number between 0 (for a single grape variety) and the natural logarithm of 5 (for equal amounts of each of the 5 varieties). The data for each vintage come from the Opus One web site. The variation in Shannon diversity is shown in the first graph, with the vintages plotted horizontally and the diversity plotted vertically.

Grape diversity through time for the Opus One wine

This graphs shows that there was a lot of variability between the first few vintages, while the winemakers worked out what wine style they were aiming for. Furthermore, from the early 1990s onwards the diversity has steadily increased. This has been partly the result of using five grape varieties, as opposed to the original three, but it is mainly a result of using greater proportions of the minor varieties. In the early years, there were vintages composed of >95% cabernet sauvignon, but over the past 10 years it has been closer to 80%. For the rest of the grapes, it has been c.7% merlot, c.6% cabernet franc, c.6% petit verdot, and c.1% malbec.

Having established that the winemakers have been moving towards a greater diversity of grape varieties in Opus One, we can now ask whether this has improved the wine quality in the eyes of the drinkers. There have, of course, been a number of retrospective tastings of the vintages of Opus One, which is getting closer to its 40th vintage. It therefore seems worthwhile to see whether the quality scores given to these wines are associated in any way with the particular mixture of grape varieties that have been included in the wine over the years.

The most complete vertical tasting that I have been able to find is that of Antonio Galloni, from 2013, which included all of the vintages from 1979–2010. Sadly, the best vintage of all has been suggested to be the 2013, which misses out. In the next graph I have plotted Galloni's quality scores (vertically) against the Shannon diversity (horizontally), with each point representing a single vintage.

Wine quality and grape diversity for the Opus One wine

The graph shows a general increase in quality score with increasing diversity, with four exceptions (as labeled in in the graph). Excluding these four vintages for the moment, a correlation analysis shows that 42% of the variation in the wine quality score is associated with the grape diversity score. That is, increasing the diversity of the grape varieties in the wine has generally improved the quality, which is presumably what the winemakers have intended.

The four exceptions are instructive. The 1980, 1984 and 1987 vintages consisted almost entirely of cabernet sauvignon (>95%), and this has obviously been a very erratic strategy in terms of wine quality (sometimes it worked and sometimes it didn't). Furthermore, the 1980 and 1984 vintages consisted solely of cabernet sauvignon and cabernet franc, with no merlot at all. On the other hand, the 2006 vintage had the highest proportion of merlot yet, at 12%, which is double the usual amount. This created a high diversity value but obviously not a high quality score from Galloni. The winemakers again tried such a high proportion of merlot for the 2011 vintage (11%), and Galloni's preliminary score for the resulting wine (not yet released) indicated that he didn't think it had worked then, either.

This pattern could, of course be unique to Antonio Galloni — I have repeatedly pointed out that wine critics rarely agree much with each other about wine quality (see How large is between-critic variation in quality scores?). However, few of the other professional commentators have conducted extensive vertical tastings of Opus One. So, by way of comparison, let's look at the opinions of a group of non-professionals.

In 2002, Bob Henry, a wine marketer from California, conducted a group tasting of the first 20 vintages of Opus One (1979–1998). Each of the 23 tasters was asked to rank their top three wines, with 3 points being assigned to the top wine, 2 points to the second wine, and 1 point to the third wine. These scores were then summed across the tasters, in order to rank the quality of the vintages. These results are compared to those of Galloni in the next graph, with each point representing one of the 20 wines.

Quality scores for vertical tastings of the Opus One wine

As you can see, only nine of the wines scored any points (ie. was a top-3 wine for any of the tasters). Most of these wines were also high-scoring wines for Galloni, and so we can treat this as a general confirmation of his scores. However, note that the 1980 vintage, which had a low grape-diversity score but still received 95 points from Galloni anyway, was not a high-scoring wine for the tasting group. This means that only the 1987 wine scored points but had a low grape-variety diversity. Indeed, the 1991 vintage was the only high-scoring wine before the introduction of malbec and petit verdot to the mix.

Conclusion

In biology (including agriculture), diversity is considered to be a Good Thing. Here, the Opus One wine seems to support this idea, as increasing diversity of grape varieties is associated with higher quality wines. Furthermore, the winemakers have been steadily increasing this diversity with each succeeding vintage. This is a strong argument for varietal diversity in wines. If nothing else, this helps explain the wine's reputation for inconsistency — poor vintages have generally arisen from reliance on too few grape varieties.

Monday, May 8, 2017

The Wine Spectator prefers modern wine styles

In some recent posts I have compared the wine-quality scores provided by different commentators. While doing the data analyses, I noticed that the scores from the Wine Spectator magazine had a particular pattern that the other scores did not — there was a time trend to the scores.

For example, in the post on Poor correlation among critics' quality scores, I compared the quality scores from five commentators over 60 vintages of the Penfolds Grange wine. There was no time trend in the scores for four of the commentators, but the Wine Spectator showed a very clear upwards trend in the scores through the vintages, as shown in the first graph. [Note: most of these wine scores were not given at the time of release, but are based on subsequent retrospective tastings.]

Wine Spectator quality scores for Penfolds Grange

For comparison, the time correlation value for the other commentators ranges from 7% to 18%, versus 58% for the Spectator. So, for the Wine Spectator more than a half of the variation in the scores is associated with time, which is not true for the other commentators.

The line in the above graph is a running average (of 9 vintages), showing that the scores rise until the early 1990s, and then remain somewhat steady after that. Indeed, it was in 1995 that the Wine Spectator named the 1990 Penfolds Grange as its wine of the year.

This intrigued me, so I looked for other long-term data from the Wine Spectator. I looked for a broad range of wine types (different styles from different regions), since the Spectator has a range of different reviewers, and I wanted to include as many of these as possible. For each wine, I wanted at least 10 scores from the period 1975-2014 (40 vintages), with some of the scores before and some after the 2000 vintage. What I came up with is shown in the table, with the time correlation indicated for each wine.

Schloss Vollrads Riesling Spätlese, Rhine
Ruffino Riserva Ducale, Chianti
Schloss Johannisberg Riesling Spätlese, Rhine
Penfolds Grange Bin 95, Barossa
E. Guigal Château d'Ampuis, Rhône
Fontanafredda Serralunga d'Alba, Barolo
J.J. Prüm Wehlener Sonnenuhr Riesling Auslese, Mosel
Viña Tondonia Reserva, Rioja
Louis Latour Corton Grand Cru, Bourgogne
Viña Tondonia Gran Reserva, Rioja
Château Latour, Bordeaux
Moet & Chandon Brut, Champagne
Château Lafite-Rothschild, Bordeaux
Henschke Hill of Grace, Barossa
Château Climens, Bordeaux
Robert Mondavi Winery Cabernet Sauvignon, Napa
Château Mouton-Rothschild, Bordeaux
Château Margaux, Bordeaux
Stag's Leap Wine Cellars Cabernet Sauvignon, Napa
Joseph Drouhin Charmes-Chambertin Grand Cru, Bourgogne
 83%
 67%
 63%
 58%
 57%
 56%
 54%
 52%
 38%
 32%
 30%
 29%
 26%
 26%
 22%
 11%
   8%
   5%
   2%
   1%

As you can see, the majority of the correlations are high — in this context, any correlation greater than about 20% is unusually high. As an illustration of what a 25% correlation looks like, in the next graph I have included the data for two of the Bordeaux chateaux for the period since 1980.

Wine Spectator quality scores for Chateaux Latour and Lafite

It turns out that I am not the first person to have noticed this pattern. In a blog post entitled Fun with wine numbers, Tom Wark looked at the Wine Spectator's point scores for several hundred California chardonnays over 18 years. He summarized the data in terms of the percentage of wines with particular scores; and I have graphed his results in the next figure.

Wine Spectator quality scores for California chardonnays

The decrease in the percentage of wines with scores <80 is the result of an editorial decision to stop publishing such scores (there are plenty of wines with high scores to write about). However, the increase in the percentage of wines with scores >90 is precisely what I have shown above for individual wines.

Why has this happened?

This leads inevitably to a consideration of what is causing this time pattern. Tom Wark commented: "I honestly don't know what to make of this. It looks like point inflation taking place ... On the other hand, it just may be that CA Chardonnays got a heck of a lot better."

The Wine Spectator itself agrees with the second option. Writing in that magazine, James Laube (Wine rating inflation) noted: "It's indisputable that wines are better now than a generation ago. Vineyard management, winery technology, winemaker skill — all have progressed. And as wines have improved, ratings as reflected by scores have risen."

However, not everyone else agrees with this idea. If we take the Henschke Hill of Grace wine listed above, for which the Spectator time correlation is 26%, the time correlations for the same period from some other commentators are: Jancis Robinson 0%, Jeremy Oliver 4%, James Suckling 5%, Huon Hooke 25%, and Robert Parker 26%. So, two of the critics agree with the Spectator about increasing quality, and three don't.

It is inevitable that Parker's wine scores went up, of course. Lots of winemakers started making "Parker wines" precisely for the purpose of getting high "Parker scores", and so it is inevitable that Parker would prefer later wines to the earlier ones — they were made especially for him. In this sense, Parker is simply the victim of his own excess — his strong wine bias has helped create a wine world that suits him well, but not necessarily anyone else, including other wine commentators. The wines styles do, however, seem to suit the palates at the Wine Spectator.

This is not really score inflation, but is instead simply another example of confirming the consequent. It is a feedback loop, in which high scores encourage wine makers to produce wine styles that will generate more high scores.

To examine the idea that the Wine Spectator's scores reflect score inflation, instead of better wines, we would need data that are independent of Robert Parker, which we do not have. However, it is worth noting that the Spectator's higher scores for recent wines occur even in vertical tastings, where all of the vintages are tasted at the same time. So, the high scores do not represent a slow creep upwards through time, but are instead a clear preference for modern wines compared to older styles.

Modern wine styles are associated with the change to riper vintages since 1990 (often attributed to global warming; see Fifty years of Bordeaux vintages), which produce "bigger" wines with higher alcohol contents. If you want to read about such things, then Dan Berger has a long series of posts in his USA-syndicated wine column (thanks to Bob Henry for pointing these out to me):
Big wines
Bored with big wines?
Bigger is better?
Bigger: Is it better?
'Big' wines
Is bigger better?
High alcohol
A troubling trend?
Another look at high alcohol content
How much alcohol should wine have?
Blind tasting and alcohol
High alcohol and diminishing wine styles

Monday, May 1, 2017

Do sales by US wine companies fit the proverbial "power law"?

The short answer is: almost.

The Power Law is used to describe phenomena where large events are rare but small ones are quite common. For example, there are few billionaires while most people make only a modest income; there are few large cities but many small towns; there are few very frequent words (such as "and", "the") but many rare words.

Mathematically, Power Laws are of interest because of what is known as "scale invariance", as well as the fact that there is no well-defined average value. You can read about this in Wikipedia.


For the rest of us, Power Laws are of interest because of their practical consequences. For example, the 80:20 Rule (or Pareto Principle) is one example of a Power Law, which says that for many events, roughly 80% of the effects come from 20% of the causes. You can also read about this in Wikipedia. For a discussion of this idea, see What is the 80/20 rule and why it will change your life. [Sometimes, it is also 90/10; for example, it is usually estimated that 90% of wine consumption in the USA is by 10% of the people.]

Power Laws are considered to be universal, and so there is no reason why they should not exist in the wine industry. One of the more obvious places that we might expect to find them is in the size of wine companies. Size might be measured by amount of wine produced or by monetary income, for example. Either way, there will be a few very big companies and lots of little ones.

So, let's look at a specific example. In its February 2017 issue (p.48), Wine Business Monthly compiled its fourteenth annual ranking of the top 30 U.S. wineries by case sales (in 2016). You can see a copy of the list in Big Wine takes over.

Wine Business Monthly notes:
Though there are now 9,091 wineries in the U.S., the WBM 30 companies represent more than 90 percent of domestic wine sold by volume. The three top wine companies by themselves represent more than half of all case sales.
This sounds like a classic case of a Power Law; and so it is worth checking this possibility.

One special case of the Power Law is known as Zipf's Law, which refers to the "size" of each event relative to it's rank order of size. This is what we are looking at here. For each wine company, the "size" is the number of cases of wine sold during 2016, and the WBM 30 companies are listed in rank order of their sizes (largest to smallest).

The standard way to evaluate the Zipf pattern is to plot the data with both axes of the graph converted to logarithms. Under these circumstances, the data should form a straight line. Here is the graph of the WBM 30 data. The three largest wineries are labeled.

A Power Law fitted to the sales by US wine companies

As you can see, almost all of the data lie roughly along a straight line, and thus do indeed fit a Power Law. That is as expected; and the Power Law is thus not proverbial in this case.

However, there is one exception — the largest company, E&J Gallo Winery, did not produce enough wine to fit into the same pattern as the other 29 wineries. Indeed, to fit the Power Law, the top-ranked company would need to have sold about 260 million cases during 2016, which is c. 3.5 times as much wine as E&J Gallo actually sold.

This is an interesting finding. The Power Law suggests that the biggest US winery (and thus the biggest wine company in the world) should dominate the US industry to a greater extent than the Gallo winery currently does. That is, having one-quarter of the US wine market is not enough! The current degree of industry dominance has been held by Gallo for at least the past quarter-century, but it seems unlikely that it has ever dominated sufficiently to fit the Power Law. Apparently, it is rather hard to dominate US wine sales in the way predicted by a simple Power Law model.

This "under-performance" by Gallo may be a good thing for the consumer, of course, since Diversity is usually a better thing than is a Power Law — the Power Law actually represents a rather extreme situation.

For comparison, the biggest-selling imported wine in the USA is Yellow Tail (from Casella Wines, in Australia), with more than 8 million cases shipped to the US per year. This would place it at no. 9 in the WBM 30 list. This will be the subject of a separate blog post.