JWE-Articles

alyses and aggregations of the ratings that wine critics and judges assign to wines are made difficult by stochastic error and biases that remain even when wines are assessed blind to price, label, capsule, and closure. Stochastic error is due to the partially random nature of ratings. Cognitive and omitted-variable biases are due to anchoring, expectation, serial position, commercial, and other factors. Differences in decanting, filtering, aeration, and temperature can also affect ratings.

This study uses gravity models to explain bilateral patterns of global wine trade since 1962. This is, to our knowledge, the first study on global wine trade covering the second wave of globalization as a whole. The results suggest that the impact of distance, common lan- guage, and common colonizer post-1945 on wine trade was lower in the 1991–2019 period than in the 1962–1990 period. We also use gravity models to explain the impact on bilat- eral wine trade patterns of similarities across countries in the mix of winegrape varieties in their vineyards. Although our models do not allow us to identify causality, the results sug- gest that countries trade more wine with each other the closer their mix of winegrape varieties.

Most of the previous research examined the demand for alcohol consumed at the off-trade (consumed at home). However, some consumers might prefer to consume alcohol on- trade (away from home) or switch between on-trade and off-trade consumption as a reac- tion to price or income change. We estimate the Quadratic Almost Ideal Demand System consisting of three broad alcohol categories, consumed on-trade and off-trade, to derive own-price, cross-price, and income elasticities. Selectivity due to the high censoring is treated, and special attention is paid to quality-adjusted price. Beer consumption is the most responsive to income as well as own price changes, while spirits are the least respon- sive. The own-price elasticity of wine is –0.66 and –1.00 at on-trade and off-trade, respec- tively. Beer is more price responsive, spirits are less price responsive, and consumption reacts weaker in the off-trade market. Own-price elasticities of demand range between –1.20 and –0.41 at the off-trade and between –1.51 and –0.63 at the on-trade alcohol mar- ket. Increasing the price of wine in one market decreases wine consumption in another one. Between the two markets, wine and spirits are complementary, and wine and beer are substitutes in both markets.

Studies suggest the inter-rater reliability of judges at wine competitions is higher than what would be expected by random chance, but lower than what is observed when experts in other fields make judgments specific to their expertise. To further contextualize the (un-) reliability of wine judging while also extending the study of fine water, we examine the inter- rater reliability of judges at an annual international competition for bottled waters. We find that the inter-rater reliability of water judging is generally better than chance and, at best, about the same as the inter-rater reliability of wine judging at some wine competitions. These results suggest that perceptible differences between fine waters exist but are less pro- nounced than those between fine wines and, also, that aesthetic standards with respect to fine waters exist but are currently less established than those for fine wines.

Much research shows that the ratings that judges assign to the same wine are uncertain. And while the ratings may be independent, research also shows that they are not identi- cally distributed. Thus, an acute difficulty in ratings-related research and in calculating consensus among judges is that each rating is one observation drawn from a latent distri- bution that is wine- and judge-specific. What can be deduced about the shape of a latent distribution from one observation? A simple maximum entropy estimator is proposed to describe the distribution of a rating observed. The estimator can express the implications of zero, one, a few blind replicates, and many observations. Several tests of the estimator show that results are consistent with the results of experiments with blind replicates and that results are more accurate than results based on observed ratings alone.

The changing legislative landscape of the U.S. wine market provides a scenario to examine the effect of regulation on the size distribution of firms. Using the variation across states and time in the sum of in-state and out-of-state adult populations between 2002–2017, and a difference in difference-style empirical model, I examine how restrictions on Direct to Consumer (DTC) sales impact the number of establishments and the employ- ment at wineries. I find that the expansion of the potential wine market by 10 M adults caused about a 3.5% increase in the number of wineries. While reduced DTC restrictions explain growth in the number of wineries, I find no effect of lessened restrictions on the number of winery employees, though there is evidence of a lagged effect. Additionally, I find that the growth of smaller wineries substantially outpaces that of larger wineries when regulations are lessened. These results suggest that regulatory barriers in particular indus- tries may allow states to maintain an artificial size distribution.

We study the effect of proximity to other wineries on the formation of new wineries and how this effect depends on winemaking history in a location. Clustering is common in the wine industry, but it also depends on other factors, such as proximity to vineyards and high-reputation wineries. Using panel data with annual observations from 1994 to 2014 on 598 zip codes within Washington State, we estimate empirical models that control for proximity to wineries, proximity to vines, proximity to income, and the presence of star wineries. We find that the elasticity of the number of wineries with respect to prox- imity to wineries outside the zip code hinges on the length of local winemaking history. For locations with 11 or more winery years prior to our sample, the elasticity is at least 0.44. The presence of elite wineries is also found to have an effect, with about 0.5 addi- tional wineries per year starting in a zip code per star winery. The effect of history suggests that policies to seed winery start-ups will help cluster formation, but only with a substan- tial critical mass of winemaking activity.

As the market for fine-wine investing matures, basic questions of portfolio strategy remain unexplored. I evaluate how adding fine wine from the superstar châteaux of Bordeaux’s Right Bank might complement the traditional focus on the five first-growths of Bordeaux’s Left Bank. Fundamentals for the Right Bank’s superstars are attractive: they produce roughly an order of magnitude less, face different production conditions, and receive equally impressive critical reviews. However, they receive far less attention than their Left Bank counterparts. To examine returns over the long run, I hand-collected 10,885 prices for eight wines from an archive of 391 Sherry-Lehmann catalogs, a New York City retailer, which began at the end of Prohibition. Using these historical price records, I compare the real returns from investing in the five Premier Cru to a port- folio that adds three superstar châteaux from the Right Bank: Ausone, Cheval Blanc, and Petrus. I find the geometric-average annual return was 6.78% in real terms from 1938 to 2017 for the joint portfolio, less than 0.01% different, but with better risk-reward as mea- sured by the Sharpe ratio. Additionally, I find the life cycle of aging is substantially differ- ent across the two Banks, which could provide further diversification benefits for the strategic investor.

Viruses and related graft-transmissible pathogens cause diseases that cost the grape indus- try billions of dollars annually if left uncontrolled. The National Clean Plant Network (NCPN), a USDA Farm Bill program, is an organization of clean plant centers that pro- duce and maintain virus-tested foundation vine stocks and distribute propagation material derived thereof to nurseries and growers to minimize the introduction of viruses and virus-like diseases into the vineyard. Foundation Plant Services (FPS) is the major NCPN-grapes center. We examined the economic impacts of public investments in FPS from 2006 to 2019. By focusing on grapevine leafroll disease, our analyses revealed a ben- efit-cost ratio ranging from 22:1 to 96:1, with a 5% and a 20% disease infection rates in commercial vineyards, respectively. A welfare analysis was consistent with grape growers and nurseries capturing most (64–98%) of the benefits from adopting clean planting material compared with winemakers and other actors in the downstream wine supply chain system. This study provided new insights into the returns to public investments in a clean plant center and documented strong financial incentives for higher adoption of clean vines derived from virus-tested stocks, while justifying continued support of NCPN centers from public and private sectors.

This paper investigates the relative importance of various attributes, including varietal, brands, and geographic origin, in explaining retail wine prices for the United States mar- ket. We use a metric based on the Shapely value, from cooperative game theory, in the context of an empirical hedonic price equation estimated using a large sample of retail wine sales for home consumption over the period 2007–2019. We find that brands alone explain more than 70% of the variation in wine prices, but geographic origin and varietals retain additional explanatory power. Furthermore, information about the geo- graphic origin appears to be a considerably more important attribute than varietals.

The outcome of the famous 1976 Judgment of Paris, a blind wine tasting of ten wines by nine French judges, brought American wines to the forefront of the wine business. A Californian wine, the 1973 Stag’s Leap Wine Cellars S.L.V. Cabernet Sauvignon, was declared the winner, surpassing four highly prized French wines (Château Mouton- Rothschild 1970, Château Montrose 1970, Château Haut-Brion 1970, and Château Léoville Las Cases 1971). We collect ratings from experts for (almost) all vintages of the same ten wines over the years 1968–2021 and find that the Stag’s Leap Cabernet Sauvignon is far from being first. We conclude that either the 1973 vintage was overrated by the experts who tasted it in 1976, or 1973 was merely an outlier in this winery.

We use data from Wine Spectator on 266,301 bottles from 12 countries sold in the United States to investigate the link between the score awarded by the guide and the price charged. The link between quality and price is positive, in line with the literature. In a deeper inspection, however, hedonic regressions show that the price premium attached to higher quality is significant only for “superstar” wines with more than 90 points (on a 50–100 scale), while prices of wines between 50 and 90 points are not statistically different from each other. Furthermore, an analysis performed through normal heteroske- dastic and quantile regression models shows that the dispersion of quality-adjusted prices is described by an asymmetric U-shaped function of the score; that is, products with the lowest and highest quality have the highest residual standard deviation. Pursuing excel- lence is a risky strategy; the average price is significantly higher only for wines that achieve top scores, and the price premium becomes more volatile.

Wine investment returns can come from overall market trends or price increases with age. Because of the short wine price histories available, market and maturation effects are difficult to separate. Consequently, researchers often obtain dramatically different estimates of investment returns. We find that data sample bias may be the hidden cause of the disparate estimates. In wine auction data, the sample bias refers to a shift in the distribution of which wines are traded as a function of their age. Such sample bias in panel data sampled across many different wine labels can distort the estimation of price increases versus age and consequently impact the estimation of market trends. This analysis shows that segmenting the analysis such that the data panels contain wine labels with similar trading characteristics can lead to a more stable estimation.

The analysis here looks at data from Bordeaux, Italy, Australia, and California. An Age-Period-Cohort (APC) analysis is applied to data panels from each region. Then the data in each region is segmented by a measure of popularity in order to reduce sampling bias. Data thus segmented is then re-analyzed to demonstrate the difference in estimating price appreciation lifecycles and market trends.

A proposal to reform the United Kingdom’s excise duty on alcohol is under consideration during 2022. The proposal would change the tax base from volume of product to volume of alcohol, which would see a fall in the tax on sparkling wine (by about one-fifth), a rise in the tax on fortified wines of 18% alcohol by volume (ABV) (by about one-sixth), and table wines with more (less) than 11.5% ABV would become dearer (cheaper). With taxes on most beers unchanged and taxes on spirits lowered slightly, the pattern of UK wine consumption and imports would alter considerably. This article draws on a global model of national alcoholic beverage markets to estimate the likely bilateral trade effects of this proposed reform to UK excise duties. It compares them with the trade effects of the United Kingdom’s first two bilateral free trade agreements (FTAs), following the post-Brexit EU–UK Trade and Cooperation Agreement, which allows Australian and New Zealand vignerons tariff-free access to the UK wine market. Those two FTAs are estimated to cause the United Kingdom to import far more wine than is lost by the proposed changes in UK excise duties.

This paper uses a difference-in-differences approach to analyze the treatment effect of a hail weather shock in a specific Swiss wine-growing region. We exploit a natural experi- ment from Switzerland’s Three Lakes wine region in 2013 and examine its impact on the country’s retail market. We find statistically significant (1%-level) effects of –22.8% and +2.8% for the volume and price of wine consumed, respectively. These effects can be inter- preted as average treatment effects, which is the difference in outcomes between treatment and control groups using a pre-post shock study methodology. Several robustness checks confirm the statistical significance of the estimated effects and the initial assumptions.