Issue link: http://winesandvines.uberflip.com/i/732978
October 2016 WINES&VINES 57 PRACTICAL WINERY & VINEYARD WINEMAKING average difference, calculated as reported alcohol minus actual alcohol, was –0.13% ABV. Wine from every country, on average, had higher actual content than was declared on the label, and the average understatement was largest among wines from countries that had higher average alcohol content (such as the United States, Australia, Spain, Chile and Argentina). Averages may be misleading These average values mask a lot of variation within countries and between red and white wines, and some of these details were the subject of further analysis in the article "Splendide Mendax: False Label Claims about the High and Rising Alcohol Content of Wine" in the Journal of Wine Economics (Alston et al., December 2015). The average figures conceal the fact that, in many in- stances, the labels overstate the alcohol content of wine, even though it is more often understated. In more than half of the observations (52,178, or 57.1% of the total), alcohol con- tent was understated. For this group, on aver- age, the actual alcohol content was 13.56%, and the reported alcohol content was 13.15%, for a discrepancy of 0.42% ABV. A discrepancy of 0.4% might not seem large relative to an actual value of 13.6% ABV, but it is much more significant compared with the typical range for wines in a particular category. For example, Napa Valley Cabernet Sauvi- gnon might be expected to have alcohol con- tent within the range of 13.5% to 14.5% ABV, and an average error of 0.4% ABV is high in the context of this range. The size of the un- derstatement was similar between red and white wines. The patterns are somewhat dif- ferent if we further divide the data in this group between New World and Old World sources; Old World wine labels understated the alcohol content to a smaller extent than New World wine labels. Labels for 29,461 wines (32.2% of the sample) erred in the opposite direction, overstating the true alcohol content by 0.01% or more. On average, the actual alco- hol content for this group was 12.9% by volume, and the reported alcohol percentage was 13.2%, with a discrepancy of 0.32%. Within this group, the size of the overstate- ment was similar between red and white wines and similar between the New World and Old World sources. More than 10% of the useful sample (9,793 observations) represent wines with reported alcohol percentage within 0.01% ABV of the actual alcohol percentage. In this category, Old World red wine had an average alcohol content of 13.0% by volume; Old World white, 12.5%; New World red, 13.6%, and New World white, 13.1%. Labeling errors might make economic sense Why do we observe these patterns of surpris- ingly large labeling errors that cannot be ex- plained as random measurement error, even allowing for commercial realities? We cannot be sure, but we can speculate. Wineries may have incentives to understate or overstate the alcohol content because they perceive a market preference for a particular range of alcohol content for a given style of wine. The wide error tolerances permit them to do this legally, and our results suggest that is what at least some of them are doing. Winemakers can manage the alcohol con- tent and other characteristics of wine to some extent and at a cost, but they cannot cheaply vary the alcohol concentration independent of other characteristics. For example, to achieve riper, more intense fruit flavors may require longer "hang time" for grapes, which would also lead to more concentrated sugar levels and higher alcohol wine. We suggest consumers may happily pay a premium for the resulting flavors yet would prefer not to have (or, at least, know about) the concomi- tant increase in alcohol content. In such a setting, it may be profitable for a winery to give the consumer both the de- sired wine characteristics and the preferred label information by under- or over-stating the true alcohol content. We base this specu- lation in part on discussions with several winemakers who have told us informally that they chose to understate the alcohol content on a particular wine label within the range of error permitted by the law. We suggest they made this choice because they believed it would be advantageous for marketing the wine to have a stated alcohol content closer to what consumers would expect to find in high-quality wine. Bottom line on wine alcohol labeling errors Our findings support the idea that winemakers AVERAGE ALCOHOL CONTENT AND REPORTING ERROR BY COUNTRY Average Alcohol Percentage Country or Region of Origin Number of Wines Tested Actual Alcohol Reported Alcohol Difference (Actual – Reported) Old World France 25,404 13.0% 12.9% –0.10% Italy 19,806 13.0% 12.9% –0.09% Spain 2,993 13.4% 13.2% –0.21% Portugal 2,321 13.0% 12.9% –0.05% Total 50,524 13.0% 12.9% –0.10% New World Argentina 1,778 13.8% 13.6% –0.24% Australia 9,617 13.7% 13.7% –0.09% Canada 4,113 12.8% 12.6% –0.13% Chile 3,744 13.7% 13.4% –0.27% New Zealand 2,125 13.2% 13.2% –0.06% South Africa 3,347 13.5% 13.4% –0.09% United States 16,184 13.9% 13.7% –0.23% Total 40,908 13.7% 13.5% –0.17% World 91,432 13.3% 13.2% –0.13% Under-reported alcohol 52,178 13.6% 13.2% –0.42% Over-reported alcohol 29,461 12.9% 13.2% 0.32% Correct alcohol % 9,793 13.1% 13.1% 0.00%