Orley Clark Ashenfelter is a professor of economics at Princeton University. Even though his academic interests are in labour economics, econometrics, and law and economics, he is also an amateur wine lover who spends a lot of time studying it and even published a research paper on wine prices. Later it occurred to him that this “amateurish” study of his would soon become his most famous discovery which upset the entire wine community for over a decade. (Ayres, 2006)
Traditionally, swishing and spitting by gurus is the only way to predict auction prices. Such decisions are entirely subject to the personal preference of the experts, and human memory tends to be biased that we are not good at dealing with complex information. Ashenfelter decided to counter this problem with an objective statistical approach: making predictions by collecting past samples and analysing possible correlations between different factors.
Wine quality is the best when the grapes are ripe, and their juice has a high concentration. In years when the summer is hot, grapes get ripe the fruit gets concentrated when the rainfall is low. Therefore, the hot and dry years tend to get the legendary vintages. It is also a fact that the quality will be even better if the preceding winter has a heavy rainfall. Armed with this knowledge, Ashenfelter came up with the following formula to predict wine prices:
Quality = 12.145 + (0.00117 * Winter rainfall) + (0.0614 * Average summer temperature) - (0.00386 * Harvest rainfall)
In other words, if you can travel around the world and collect the data for the three variables in the above equation, you will be able to compute the approximate price of the wine even without opening the bottle and tasting it.
It goes without saying that the experts did not like the formula. Britain’s Wine magazine scorned that the formula’s plain silliness invites disrespect, while famous American expert Robert Parker called Ashenfelter “an absolute total sham” who is rather like a critic who evaluates a film without seeing the actual performance. Their reactions are understandable. If a formula could do such an amazing job to assess a wine’s quality, then there will be no more need for pundits who possessed years of experience and who had drunk lots of wine.
However, Ashenfelter’s formula had done so well that it was like punching the experts in their faces. For example, Parker rates the 1986’s as “very good and sometimes exceptional”. Peter A. Sichel, the author of the influential Bordeaux Vintage and Market Report, said the 1986’s have “elegance and classic Bordeaux structure”. New York stores, brimming with the vintage, are pricing the wines in the same range as the much-praised 1985’s.
In contrast, according to Ashenfelter’s system, below-average growing season temperatures and above-average harvest rainfall doom the 1986 Bordeaux to mediocrity. It predicted that the 1986’s would be the worst vintage of the 1980’s, and no better than the unmemorable 1974’s or 1969’s. The formula turned out to be correct, and many experts had no choice but to revise their ratings.
Soon after this, Ashenfelter’s formula made another important prediction for the 1989 vintage, one which would cement its place in the wine industry if it is successful. At the time of its prediction, the wines were just three months in the cask and not judged by the critics at all, but Ashenfelter had already come out with a verdict. The hottest growing season in recent years and a dry harvest weather in 1989, according to the system, would guarantee that the quality to be stunningly good. Adjusted for age, it predicted that these wines would eventually sell for a substantial premium over the great 1961 vintage.
As usual, the professionals were not amused by Ashenfelter’s presumption, but they could not argue with the facts. The 1989 vintage was so great that it sold at a price almost double of that of the 1986’s, one which they previously praised a lot. Immediately following this, Ashenfelter predicted that the 1990’s would be a classic, even better than the 1989’s, and it turned out to be correct again.
As Nobel Prize winner Daniel Kahneman (2011) observed, in fields of high unpredictability, such as professional sports and the stock market, expert performances are often inferior to algorithms, because humans are prone to different logical biases. By employing a statistical approach, one is not only able to remove the undesirable effects of mental distortions but also to align oneself with the objective facts. It is why he wrote, “Whenever we can replace human judgement by a formula, we should at least consider it.”
REFERENCE:
Ayres, I. (2006). Super Crunchers: Why thinking-by-numbers is the new way to be smart. New York, NY: Random House.
Kahneman, D. (2011). Thinking, Fast and Slow. New York, NY: Farrar, Straus and Giroux.
This is from a very old study (2006) from a single older vintage (1989) — a classic example of having to scrutinize what you read on the internet!
ReplyDeleteHere's my formula for wine prices: 30% place of origin X 30% producer reputation X 20% wine writer bullshit X 20% producer bullshit = wine price.
I dare anyone to prove this wrong, tho I welcome any alterations to the percentages.