Like previous editions, Goldstein and Herschkowitsch’s The Wine Trials 2010 would be a valuable addition to any casual or serious wine drinker’s bookshelf. The pithy descriptions of the 150 best wines under $15 provide information about each wine that, quite remarkably, is not collated elsewhere in a convenient usable format.
If such descriptions of good value wines were all that this book accomplished, the book would probably not merit a review in this journal. However, this book also has considerable interest to economists. The source of this interest is two-fold. First, this book convincingly describes how consumers’ lack of information about individual wines may generate social welfare losses and argues that blind tasting is the best way to eliminate this information problem. Second, the book follows its own advice and reports data on the best wines collected using modern methods that should be familiar to any economist doing empirical work today. The ultimate result of Goldstein and Herschkowitsch’s efforts may be no less than a fundamental shift in the way the wine market operates to an environment in which value, as measured by taste rather than perception, is paramount.
In the first section of the book, the authors eloquently detail how today’s wine industry actively encourages consumers to pay far above marginal production costs to drink wines that simply do not taste any better on average than inexpensive wines to most consumers. They provide evidence that the information consumers have about wines prior to purchasing them is largely controlled by the wine industry. More expensive wines are rated higher by industry publications like Wine Spectator even when blind taste tests reveal no significant relationship between price and taste for everyday consumers (Goldstein et al., 2008). Even self-proclaimed wine “experts” do not find the more expensive wines to taste much better. The authors exposit that much of the reported enjoyment consumers derive from drinking wine has nothing to do with its taste but rather the perception that they are drinking a good wine. Because they are stuck in an environment that rewards flash and marketing over taste, even small producers have a disincentive to produce wines with the best taste, thereby perpetuating the problem. Ultimately, the book’s first section is a critique of the wisdom of drawing welfare conclusions from revealed preference data in an environment with asymmetric information. The authors’ critique of the wine industry should apply equally to other consumer products that feature many closely substitutable varieties about which information is scarce to the consumer.
The wine reviews themselves, which make up the bulk of the second part of the book, are written by the authors based on reports of blind taste testers. This amounts to a one-page report on each of the 150 highest ranked wines one by one including useful suggestions for food pairings. Most interesting to a general (wine-neutral) economist audience is how the underlying data points are collected. The authors randomized 450 wines into numerous blind tastings and selected the 150 best for the book based on the resulting data using reports filled out by testers. The fact that the resulting data set captures individuals’ preferences for only the taste
component of the wine drinking experience across wines leads to some intriguing questions.
Subject to the technological constraints of the human palate, the authors do a commendable job of data collection. The blind tastings organized by the book’s authors successfully achieved random allocation of wine attributes not observed by tasters (like price) across the tasting population. This facilitated successful isolation of preference relations over taste not conflated with other attributes of the wines. Assuming that individuals’ wine preferences are not perfectly idiosyncratic, something that could be tested with the blind tasting data, the same preference relations also exist on average out of sample, meaning that the book contains valuable welfare-enhancing information for the rest of us. Elsewhere in the economics literature, the closest that one comes to this sort of exercise are the speed-dating experiments conducted by Fisman et al. (2006, 2008). However, the Wine Trials authors are in some sense more ambitious. Rather than simply report on the attributes of wines that people prefer, they come to normative conclusions about inefficiencies in the wine market that crucially depend on the underlying model of behavior that generates the data.
Two mechanisms come to mind as potentially important in rationalizing the claim that pricier wines are not more enjoyable. Kamenica (2008) formalizes a model in which the choice set contains information about the (unobserved) quality of each choice. In this context, his model would argue that the authors are essentially correct, that people do not drink more expensive, heavily marketed, or cute-label wines because they actually like them better but because in the face of costly information acquisition, they have some evidence that these wines are popular. The fact that the popularity is a choice of the wine industry itself is either lost on people or an accepted second-best outcome for risk-averse consumers because it at least indicates that the wine they are buying is not God-awful. In this environment, the $14.95+tax cost of this book is a small price to pay to improve one’s wine selections and the Wine Trials series clearly has tremendous social value for it breaks the wine industry’s near monopoly on information.
A competing explanation is that the utility most people derive from drinking a glass of wine is actually a function of both the taste and the prestige, which is increasing in price, marketing, cuteness of label or experts’ or peers’ ratings. This is akin to the “warm glow” effect demonstrated to be an important component of charitable giving (Karlan and List, 2007). To be clear, suppose that the wine drinking component of utility has the production function V(T,R(p)) for Taste and Prestige, which itself is increasing in wine price p. Therefore, we can express utility as U(V(T,R(p)),y-p). Based on evidence in Goldstein et al. (2008), assume that T is not a function of p. Given some continuous distribution of income y in the consumer population, wine producers will sell wines at all points in the optimized price distribution. As is argued in the extensive hedonic literature going back to Rosen (1974), consumer heterogeneity means that the equilibrium price function is likely to be steeper than the marginal willingness to pay function for better “quality” wine, as measured by V or prestige. That is, pricier wines are “better” in the sense that they deliver a higher R, but not as much better as the price difference implies because of positive sorting on income.
As a benchmark, assume that producers are perfectly competitive and all have access to the same technology. If they are pricing at marginal cost, it must be that the inputs into producing expensive wines are more costly. Evidence reported in the book demonstrates that the marketing costs of prestigious wine producers are indeed much greater than those of producers of table wines. But there is an inherent inefficiency here in that the equilibrium price-prestige relationship is pinned down by consumer preferences. Therefore, achieving high prices given perfect competition necessitates high marketing costs assuming wine production costs vary little with prestige. One way of interpreting the authors’ lamentations about the current state of the wine market is thus not that people have preferences over prestige per se, but that prestige is increasing in price and thus is also increasing in marketing. If a superior rating in independent publications such as The Wine Trials establishes itself as an alternative marker of prestige, then this relationship will be broken and preferences reduce to become a function only of taste. However, if prestige is truly a structural function of price then there is little scope for welfare improvements. While it may be theoretically possible to engineer non-distortionary lump-sum rebates back from producers to high price paying consumers, this is probably not practical.1
My first explanation for the current state of the wine market leads the question as to why a market for better information heretofore has not been well developed. In some sense, one may argue that The Wine Trials is a solution to such a market failure. Since information about wines’ quality is a public good, there might have been under-provision by the market. If the second model is more accurate, one may ask why wine producers do not establish their prestige by using means other than price, allowing them to undercut their competitors. One answer is that there is no incentive to shift this equilibrium because they would all make 0 profits either way. The only difference would be that in the current environment the rents go to promoters and in the alternative environment most of those rents would go to consumers. If it is costly to shift the equilibrium, no producer would find it to be a privately worthwhile investment, even if it may be socially worthwhile. Once again, in this environment The Wine Trials could be the conduit through which the primary marker of prestige changes.
In the beginning of this review, I stated that The Wine Trials would be a valuable resource for the expert wine drinker as well as the casual one. My claim is in contrast to Eric Asimov’s criticism in his “The Pour” blog posting from April 22, 2008, a piece also discussed in the book. Asimov’s central argument is essentially that wine aficionados have enough experience that they know what they are doing, and by revealed preference they therefore must be doing things right. His claim is that the wedge between experts’ opinions and the blind tasting results is attributable to the setting. While this argument may be correct, meaning that non-taste attributes are important, I submit that it is still important for people who claim to be wine experts to actually think deeply about the attributes of the wine drinking experience that lead to their enjoyment and to non-experts’ enjoyment. Otherwise, they cannot truly be experts. The authors’ introduction to The Wine Trials provokes such deep thinking and self-reflection. It may additionally be the case that while experts do not find more expensive wines much tastier on average, they know wines well enough to only drink the expensive wines that they themselves find tastier. Data from The Wine Trials may provide valuable new insights into this selection process.
The accuracy of the authors’ claims of inefficiency in the wine market does ultimately rest on whether price enters directly into people’s preferences, or whether price is just an indicator of status or a perceived indicator of taste. If price enters directly into utility, one can legitimately argue over the authors’ assertion that people would be better off if it did not. The debate over endogenous preferences is well beyond the scope of this review, but it is a serious debate with valid points on both sides. It may be true that we would all be better off if we did not care directly about the price of the wines we drink, but such a claim is probably not empirically testable. However, it does seem that it would be feasible to empirically distinguish between the preference relations over price, information and prestige in wine tastings for future editions. In future tastings, Goldstein and Herschkowitsch may wish to additionally randomize over listed prices and another measure of prestige to distinguish whether people enjoy drinking higher price or higher prestige wines more, independent of taste, ultimately recovering the V function. In addition, it would be instructive to have versions of the trials in which there is full disclosure about the wines being tasted to serve as a baseline control set. In any case, The Wine Trials is an interesting, provocative and eminently useful book that through its mere existence has the potential to fundamentally improve the functioning of the wine market.
Asimov, E. (2008). A closer look at the ‘Wine Trials’. The Pour Blog, April 22, The New York Times.
Goldstein, R., Almenberg, J., Dreber, A., Herschkowitsch, A. and Katz, J. (2008). Do more expensive wines taste better? Evidence from a large sample of blind tastings. Journal of Wine Economics, 3(1), 1–9.
Fisman, R., Iyengar, S., Kamenica, E. and Simonson, I. (2006). Gender differences in mate selection: evidence from a speed dating experiment. Quarterly Journal of Economics, 121, 673–697.
Fisman, R., Iyengar, S., Kamenica, E. and Simonson, I. (2008). Racial Preferences in Dating. Review of Economic Studies, 75(1), 117–132.
Kamenica, E. (2008). Contextual inference in markets: on the informational content of product lines. American Economic Review, 98(5), 2127–2149.
Karlan, D. and List, J.A. (2007). Does price matter in charitable giving? Evidence from a large-scale natural field experiment. American Economic Review, 97(5), 1774–1793.
Rosen, S. (1974). Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition. Journal of Political Economy, 82, 34–55.
1) Alcohol taxation is the primary mechanism by which such rebating already occurs, though unlike lump-sum transfers it is likely to involve deadweight costs.