WEP – Wine Economics and Policy Just Accepted Manuscript 1 Just accepted 1 2 3 4 U.S. Wine Consumer Interest in Wine Ingredient and Nutritional Information 5 Terry M. Lease1 (corresponding author), Deirdre Sommerlad-Rogers2 6 7 8 9 1 College of Agriculture, Food and Environmental Science, California Polytechnic State University 10 1 Grand Avenue, San Luis Obispo, CA 93407 USA, E-mail: lease@calpoly.edu 11 2 College of Agriculture, Food and Environmental Science, California Polytechnic State University 12 1 Grand Avenue, San Luis Obispo, CA 93407 USA, E-mail: dsommerl@calpoly.edu 13 14 15 16 Correspondence concerning this article should be addressed to Terry M. Lease, College of 17 Agriculture, Food and Environmental Science, California Polytechnic State University 1 Grand 18 Avenue, San Luis Obispo, CA 93407 USA, E-mail: lease@calpoly.edu 19 20 21 22 This article has been accepted for publication and undergone full peer review but has not been through 23 the copyediting, typesetting, pagination and proofreading process, which may lead to differences 24 between this version and the Version of Record. 25 26 Please cite this article as: 27 28 Lease T.M., Sommerlad-Rogers D. (2022), U.S. Wine Consumer Interest in Wine Ingredient and 29 Nutritional Information, Wine Economics and Policy, Just Accepted. 30 DOI: 10.36253/wep-12577 31 32 33 34 35 mailto:lease@calpoly.edu mailto:dsommerl@calpoly.edu mailto:lease@calpoly.edu WEP – Wine Economics and Policy Just Accepted Manuscript 2 Abstract 36 37 In the United States, labelling for wine containing at least 7% alcohol by volume is regulated by the 38 Tax and Trade Bureau, which does not require wine labels to include ingredient or nutrition labelling, 39 except for added sulfites. With the European Union moving toward mandatory disclosure of nutrition 40 and ingredient information for wine, one may expect the level of debate in the U.S. to increase. We 41 conducted an online survey of consumers in the U.S. who are at least 21 years old (legal drinking age 42 in the U.S.) and consume wine at least once every two or three months to determine their interest in 43 wineries disclosing ingredient and nutrition information for wine. We asked about the importance of 44 ingredient information when deciding which wine to purchase and when determining willingness to 45 pay, and we asked about the importance of nutrition information when deciding which wine to 46 purchase. We separately regressed three dependent variables against Wine Consumption (frequency), 47 Price, Physical Activity, Diet, Wine Knowledge, Age, Income, and Education. Overall, respondents 48 indicated that having ingredient and nutrition information was only somewhat important, with mean 49 responses 3.04 on a 5-point scale (1 = Not Important, 5 = Very Important) for ingredient information 50 when choosing a wine, 3.01 for ingredient information when determining willingness to pay, and 2.48 51 for nutrition information when choosing a wine. The factor with the greatest impact on interest in 52 ingredient information was Price, with consumers who buy a higher-end wine at least monthly having 53 a higher level of interest, followed by Diet, with consumers with a healthy diet having a higher interest 54 in ingredient information, and Age, with older consumers having less interest in ingredient 55 information. Price, Diet, and Age also had the greatest impact regarding interest in nutrition 56 information, following the same direction but with Age being the most significant. 57 58 Keywords: ingredient and nutrition information, U.S. wine consumers 59 60 61 1. INTRODUCTION 62 63 In the United States (U.S.), the labelling requirements for prepared or processed food products are 64 regulated by the Food and Drug Administration (FDA). The FDA generally requires food 65 manufacturers to list all ingredients of a food product on the label and requires most foods to bear 66 nutrition labelling. However, labelling for wine containing at least 7% alcohol by volume is not 67 covered by FDA regulations and is instead regulated by the Tax and Trade Bureau (TTB). The TTB 68 does not require wine to bear nutrition labelling, and the only ingredient requiring listing is added 69 WEP – Wine Economics and Policy Just Accepted Manuscript 3 sulfites. Alcohol by volume must be listed, but alcohol is neither an added ingredient nor a nutrition 70 category. 71 In Europe, the European Commission has rejected self-regulation proposals from the beverage 72 alcohol industry, and the European Union (EU) is moving forward with a proposal for mandatory 73 ingredients and nutrition labelling on alcoholic beverages. The European Commission’s proposal is 74 part of the “Europe’s Beating Cancer Plan” adopted on February 3, 2021, with a 2021 – 2025 75 timeframe for the alcohol-related initiative. The debate on the issue in the U.S. has been building over 76 the past few years, but as of now the TTB has not indicated interest in expanding wine labelling 77 requirements for ingredients or nutrition information. 78 Public opinion on the topic in the U.S. is unclear. Forbes [1] quotes John Gillespie of the market 79 research company Wine Opinions as saying, “I can say from a number of consumer research projects 80 in the past, ‘involved wine drinkers’—those who account for the greatest percentage of wine 81 purchases—are usually interested in having more information and detail, especially as concerns 82 health or wellbeing issues. I do think that would have an impact on how wineries respond to the 83 possibility of mandatory ingredient labeling.” 84 However, in a survey conducted by the Wine Market Council (WMC) in May 2020 [2], 41% of 85 regular wine drinkers said they rarely want to know nutritional information or the ingredient list, and 86 only 21% said they always want to know. When asked to choose the top five categories of information 87 they wanted to see on a wine label, only 4% put nutritional information and ingredient list as most 88 important, and 81% did not include it in the top five. Interest in ingredients and nutrition information 89 was positively correlated with level of education and negatively correlated with age. Core wine 90 drinkers expressed more interest than Marginal wine drinkers in this information, but they expressed 91 more interest in most forms of information, and nutrition information and ingredient list were not 92 highly ranked in the list of types of information they want. Moreover, Core wine drinkers valued 93 having additional information for the sake of knowing more about the wine and did not place much 94 value on having information as an aid in making wine purchasing decisions. 95 With the EU moving toward mandatory disclosure of nutrition and ingredient information for wine, 96 one may expect the level of debate in the U.S. to increase, as illustrated by a pair of posts on the wine-97 searcher.com website in which wine writer W. Blake Gray argued that the U.S. should follow the EU 98 on this issue [3] while wine maker Adam Lee responded with a list of challenges such regulation 99 would create and reasons why the labels could potentially cause consumer confusion [4]. This study 100 seeks to contribute additional information to the debate in the U.S. on adding the ingredient list and 101 nutrition information to required disclosure for wine and to add insight into the value of such 102 disclosure from a policy perspective. 103 WEP – Wine Economics and Policy Just Accepted Manuscript 4 104 2. LITERATURE REVIEW 105 106 Much of the academic literature related to nutrition information and ingredient list disclosure for wine 107 has centered on Europe, which is not surprising since the EU is closest to making such disclosure 108 mandatory. Bazzani, et al., [5] conducted an online survey of Italian red wine drinkers that included 109 questions on consumer attitudes toward wine and health-related aspects and a choice experiment 110 using attributes that are often associated with more natural and healthier foods. They found that health 111 consciousness is an important driver in the use of wine labels, but they did not specifically include 112 nutrition information or an ingredient list on the label. 113 Multiple studies show that the usefulness and value of nutrition and ingredient information vary 114 across countries. Employing a discreet choice experiment with representative samples of wine 115 consumers from Germany, Italy, and Australia, Pabst, et al., [6] found that consumers across all three 116 countries had a significant positive utility for detailed nutrition information. Ingredient information, 117 on the other hand, received a positive utility only in Italy, and a short ingredient list was preferred to 118 a long ingredient list. Grunert, et al., [7] utilized an online survey in Denmark, Germany, Netherlands, 119 Poland, Spain, and UK to examine consumer wants and use of ingredient and nutrition information 120 from a range of non-label sources. Information wants and use varied between the countries, was 121 highest in Spain, and was lowest in Denmark. Product involvement was a stronger predictor of 122 information wants than health interest. The effect of product knowledge was lower still and decreased 123 with more product knowledge. Previous ingredient knowledge led to lower ingredient information 124 wants, while previous nutrition information knowledge led to higher nutrition information wants. The 125 strongest predictor of information use was information wants. 126 Annunziata, et al., [8] conducted a survey in France, Spain, Italy, and the U.S. and found that interest 127 in receiving additional information on wine labels (e.g., about ingredients or nutrition information) 128 differed significantly between consumer groups based on the consumer’s country. Among 129 respondents in the U.S. panel, 40% said they seldom change a habit because of the nutritional label, 130 and the mean response for the question “I find it difficult to understand nutritional labels” was 3.1 on 131 a 5-point scale (5 = strongly agree). Only 25% said they always read the front label on a bottle of 132 wine, and only 18% said they always read the back label. Still, the mean interest in having nutritional 133 information was 3.6, and the part-wise utility value for nutritional information was higher than for 134 price, health warnings, or units in bottle and units not to exceed. 135 Another consistent result in research studies is that the usefulness and value of nutrition and 136 ingredient information are not identical across consumer segments. Escandon-Barbosa and Rialp-137 WEP – Wine Economics and Policy Just Accepted Manuscript 5 Criado [9] used eye tracking to study a sample of 114 individuals in a simulated supermarket with 138 more than 100 wines at a university in Columbia, focusing on purchase intention, related to wine 139 label information on denomination of origin, nutritional information, and health warnings. Expert 140 wine consumers used all three pieces of information to make a purchase decision. Non-expert wine 141 consumers, by contrast, made much less use of this information to make a purchase decision and 142 tended to focus on the origin information and health warnings and not make use of nutritional 143 information. The intent to purchase wine increased with the use of all three pieces of information for 144 both men and women. However, the effect was stronger for men. Women and men processed the 145 information differently, and the mean time to make a purchase decision was less than half as much 146 for men as it was for women. 147 Annunziata, et al., [8] found that interest in receiving additional information on wine labels differed 148 significantly between consumer groups based on the consumer’s socio-demographic variables, wine 149 consumption habits, attitudes towards nutritional information in general, and the degree of 150 involvement with wine. In a survey of Italian wine consumers, Annunziata, et al., [10] found that 151 consumers who already have better knowledge of wine nutritional properties and a greater awareness 152 of the links between wine and health preferred a more detailed nutritional label than other consumers. 153 Those who generally find it more difficult to understand nutritional labels either show higher interest 154 in health warnings or prefer the specification of the number of glasses not to exceed and did not value 155 more detailed information. Pabst, et al., [11] assessed consumers’ reactions to new back-label 156 information on ingredient and nutrition labelling in three focus groups with a total of twenty-one 157 wine-involved participants in three different cities in Germany. Of those participants who looked at 158 the back label (81%), almost two-thirds said they did not detect the nutrition or ingredient listing. 159 Pabst, et al., [12] conducted an online survey of German wine producers to examine producers’ 160 expectations about consumer reactions to new label information, the consequences of mandatory 161 labelling on production processes, and relative competitive advantages for different producer sizes. 162 They found that producers expect the labelling regulations to create consumer confusion and 163 uncertainty; weaken wine's image as a natural product; and increase costs due to changes in 164 oenological practices, the increased need for laboratory analyses, and more challenging labelling 165 processes. Producers believe the regulations will create opportunities for wineries to focus on clean 166 labelling strategies by completely avoiding additives that require labelling and that large wineries 167 will be better able to react to the regulations. 168 Producers’ concern for how consumers will react to the new labelling requirements is not unfounded. 169 Pabst, et al., [9] found that focus group participants who recognized the nutrition labelling and 170 ingredient list initially reacted to this information with insecurity, confusion, and incomprehension. 171 WEP – Wine Economics and Policy Just Accepted Manuscript 6 Pabst, et al., [6] found that presenting negative media information resulted in subjects in all three 172 countries surveyed significantly increasing their rating of importance of ingredients while also 173 increasing their preference for clean labelled products without ingredients. Further, a significantly 174 higher share of consumers in Germany and Italy prefer not to buy any wine. The effect of reading 175 positive media information on consumers’ wine choice is significantly lower than that of reading 176 negative information. 177 Hayward, et al., [13] studied the influence an ingredient list had on the sensory perception of red 178 wines from Nova Scotia. In this study, participants used attributes associated with liking the wine 179 more often when the ingredient list was shorter and familiar. Hayward and McSweeney [14] studied 180 the influence calorie information had on the sensory perception of rosé wines from Nova Scotia and 181 found that the calorie information did not influence consumers’ sensory perception. 182 One factor that is still undecided in the E.U. is the format of the disclosure, with producers generally 183 hoping that technology-enabled disclosure will be allowed in lieu of labelling on the bottle. Vecchio, 184 et al., [15] conducted an incentive compatible artefactual field experiment that indicated that Italian 185 wine consumers most prefer to have nutritional information presented in a panel and least prefer 186 having only a link to a website that contains the information. Grunert, et al., [7] found that the level 187 of both information wants (for ingredient and nutrition information) and information use was higher 188 for websites (product, public, and health) than for advertising, apps, or in-store sources. 189 Robinson, et al., [16] conducted a rapid systematic review and meta-analysis of eighteen studies to 190 assess consumer knowledge of energy content (calories) of alcoholic drinks, public support for energy 191 labeling, and effect of such labeling on consumer behavior. They found consistent evidence that 192 consumers tend to overestimate the number of calories in an alcoholic drink and that people are more 193 likely to support than oppose energy labeling of alcoholic drinks, but there was a high degree of 194 heterogeneity. (Two thirds of the studies used for this analysis examined nutrition information that 195 included calories, and one third looked specifically at calorie information disclosure.) The authors 196 concluded that the studies they included suggest that energy labeling did not affect consumer behavior 197 but that the overall quality of the evidence supporting that conclusion was very low. Generally, the 198 authors found that the use of self-reported information and lack of real-world settings resulted in most 199 (72%) of the studies they reviewed provided low evidential value with high levels of uncertainty. 200 Overall, the body of work shows there is inconsistency across consumers in the important of both 201 nutrition and ingredient labeling. This includes how they might use it and how much content on the 202 labels they would find important. Additionally, much of the research had been conducted in Europe. 203 The current project seeks to continue to fill the gap in how important information is to consumers, 204 targeting a U.S. sample. 205 WEP – Wine Economics and Policy Just Accepted Manuscript 7 206 3. MATERIAL AND METHODS 207 208 We conducted an online survey of consumers in the U.S. who are at least 21 years old (legal drinking 209 age in the U.S.) and either consume or purchase wine at least once every two or three months. A 210 professional panel recruitment agency recruited respondents across the U.S. using its internal 211 recruiting platform. Respondents who did not finish the questionnaire, including respondents who 212 failed a quality control check embedded in the survey, were eliminated. We received 331 completed 213 surveys. Thirteen respondents were rejected based on a speed test (completing the survey in less than 214 half the median time in a soft launch of the survey), and we obtained 318 useable responses, with an 215 average completion time of 10 minutes, 38 seconds. See Table 1 for demographic information on our 216 sample. 217 To verify that our respondent set is representative of regular wine drinkers in the U.S., we compared 218 it to the Wine Market Council’s (WMC) U.S. Wine Consumer Segmentation study, one of the most 219 thorough such studies in the industry. Comparing our respondent set to wine drinkers in the 20191 220 U.S. Wine Consumer Segmentation study [17], our set skews older. Our respondents have an average 221 age of 53.8 compared to 48.2 for the WMC study, and we have a lower percentage of respondents in 222 each 10-year age group (21 – 29, 30 – 39, etc.) below 60. Females are overrepresented in our 223 respondent set, 66% compared to 54% in the WMC study2. In terms of educational attainment, our 224 respondent set is highly comparable to the WMC study, with the same proportion of respondents who 225 did not earn any degree beyond high school (44%) and the same proportion with postgraduate work 226 or degree (20%). We have slightly more respondents with a technical or two-year degree (13 v. 11%) 227 and slightly less with a four-year degree (23% v. 25%). Respondents who identified as non-Hispanic 228 Caucasian are overrepresented (79% v. 67%). Blacks and African Americans are almost equally 229 represented in our study (10% v. 11%), but we have proportionately about half as many Hispanics 230 (7% v. 14%), Asians (2% v. 4%), and respondents identifying with another designation (2% v. 5%). 231 232 Table 1. Demographic Characteristics of the Sample 233 Frequency Percent Gender Female 208 65.409 Male 110 34.591 1 The most recent study available as of this writing 2 None of the 318 respondents either identified as non-binary or preferred not to indicate a gender. WEP – Wine Economics and Policy Just Accepted Manuscript 8 Race Caucasian/Non-Hispanic 252 79.245 Hispanic or Latino 21 6.604 Black or African American 31 9.748 Asian 7 2.201 Mixed Race 4 1.258 Other 3 0.943 Marital Status Married, in an official civil union, or in a registered domestic partnership 148 46.541 Living with a partner 31 9.748 Single, never married 64 20.126 Separated or divorced 51 16.038 Widowed 24 7.547 Income < $35,000 106 33.333 $35,000 – $49,999 56 17.610 $50,000 – $74,999 48 15.094 $75,000 – $99,999 44 13.836 $100,000 – $149,999 37 11.635 $150,000 or above 21 6.604 Prefer not to state 6 1.887 Education High school graduate or less 68 21.384 Some college 71 22.327 Completed technical/2-year degree 42 13.208 Completed 4-year degree 74 23.270 Some graduate school 11 3.459 Completed graduate Degree (e.g., MA, MS) 43 13.522 Completed terminal degree (e.g., PhD, MD, JD) 9 2.830 N = 318 Note: percentages may not add to 100% due to rounding 234 235 Geographically, the northeast U.S. is slightly underrepresented compared to the WMC study (17% v. 236 20%), with the difference divided nearly equally as overrepresentation of the mid-west, south, and 237 WEP – Wine Economics and Policy Just Accepted Manuscript 9 west regions. However, our sample set closely mirrors the distribution of the entire U.S. population, 238 with less than one percentage point difference in any region (Table 2). 239 240 Table 2. Sample and U.S. Population Distribution by Region 241 Sample U.S. Frequency Percent Frequency Percent Midwest 66 20.755 68,329,004 20.817 Northeast 53 16.667 55,982,803 17.055 South 120 37.736 125,580,448 38.259 West 79 24.843 78,347,268 23.869 N = 318 Note: percentages may not add to 100% due to rounding 242 Table 3 reports the frequency of wine consumption for our sample. The 318 usable responses include 243 six whom the Wine Market Council would not consider a wine drinker, since four drink wine less 244 than every 2 – 3 months and two never drink wine. We initially included these in the respondent set 245 because they purchase wine regularly, at least once every 2 – 3 months. Because the number of 246 respondents in this category was too small to analyze as a sub-group, we excluded them from further 247 analysis. None of the six purchased wine at a high level of frequency, five only once every 2 – 3 248 months and one 2 – 3 times per month. 249 250 Table 3. Wine Consumption Frequency 251 Frequency Percent Every day 36 11.321 Not every day but more often than once a week 79 24.843 Once a week 69 21.698 2-3 times a month 74 23.270 Once every 2-3 months 54 16.981 Less than once every 2-3 months 4* 1.258 Never 2* 0.629 N = 318 (* excluded from further analysis) 252 WEP – Wine Economics and Policy Just Accepted Manuscript 10 The WMC defines Core wine drinkers as those who report drinking wine at least once per week and 253 Marginal wine drinkers as those who drink wine less often (but at least once every 2 – 3 months) and 254 say that they like wine. The WMC definition of Total Wine Drinkers also includes those who report 255 drinking wine at least once every 2 – 3 months but say that they do not like wine. Of the 312 256 respondents in our sample whom the WMC would classify as Total Wine Drinkers, 184 (59%) are 257 Core wine drinkers, 113 (36%) are Marginal, and 15 (5%) did not report liking wine3. We conducted 258 a Pearson's chi-squared test to compare our respondent set with the WMC 2019 segmentation study 259 regarding the proportion of Core wine drinkers, Marginal wine drinkers, and others in the Total Wine 260 Drinker category. Base on Χ2 (2) = 4.655 we rejected the null hypothesis that the two groups are 261 different at p = 0.098. For the remainder of our analysis, we define Core and Marginal wine drinkers 262 based only on the frequency of wine consumption and disregard whether they report liking wine. 263 To determine the importance of having information about a wine’s ingredients, we asked respondents 264 to indicate the level of importance of knowing the wine’s ingredients when purchasing wine for each 265 of five different occasions: giving wine as a gift; bringing wine to a large gathering; bringing wine to 266 a small dinner with friends; buying wine for a special occasion at home; and buying wine simply to 267 drink at home. Using a 5-point Likert scale, respondents indicated whether, for each occasion, 268 knowing a wine’s ingredients is (1) Not Important, (2) Slightly Important, (3) Somewhat Important, 269 (4) Important, or (5) Very Important. 270 We first asked about the importance of having information about a wine’s ingredients when deciding 271 which wine to purchase. Then we asked about the importance of having information about a wine’s 272 ingredients when deciding how much the respondent would be willing to pay for the wine. We 273 calculated the mean response for each respondent across all five occasions for each question as the 274 dependent variables Ing-Choice and Ing-Pay, respectively. 275 To measure the importance of having nutritional information about a wine, we asked respondents to 276 indicate the level of importance on the same 5-point Likert scale of thirteen nutrition elements4 when 277 deciding which wine to purchase, and we calculated the mean response for each respondent as the 278 dependent variable Nutrition. 279 For each dependent variable we ran a separate regression using the following independent variables 280 that had some significance during preliminary bivariate analysis: 281 • Wine Consumption: We divided respondents into (1) Core or (2) Marginal wine drinker as 282 defined earlier. 283 3 One respondent reported being too new to wine to have a decided yet whether he or she would claim to like wine. We included that respondent in the third group for the Chi-square analysis. 4 The thirteen nutrition elements were Calories, Total Fat, Cholesterol, Sodium, Potassium, Total Carbohydrates, Sugar, Protein, Calcium, Iron, Vitamin B-6, Magnesium, and Phosphorus. WEP – Wine Economics and Policy Just Accepted Manuscript 11 • Price: We categorized respondents based on the highest price level at which they purchase 284 wine at least monthly: (1) Do not purchase wine at least once a month, (2) Purchase wine at least 285 once a month costing under $15 per 750 ml bottle, (3) Purchase wine at least once a month costing 286 $15 – $24.99 per 750 ml bottle, (4) Purchase wine at least once a month costing $25 or more per 750 287 ml bottle. 288 • Physical Activity: We asked respondents whether they regard themselves as (1) Much less 289 active, (2) Less active, (3) About the same, (4) More active, or (5) Much more active compared to 290 others their age. This was dummy coded into healthy (4 or 5 = 1) and all others (0) to compare those 291 who were intentionally engaging in a healthy lifestyle to everyone else. 292 • Diet: We asked respondents whether they would describe their diet as (1) Very unhealthy, (2) 293 Unhealthy, (3) Neutral, (4) Healthy, or (5) Very healthy. This was dummy coded into a healthy diet 294 (4 or 5 = 1) and all others (0) to compare those who were intentionally engaging in a healthy lifestyle 295 to everyone else. 296 • Wine Knowledge: We asked respondents to describe their level of wine knowledge and 297 familiarity as (1) Almost none at all, (2) Low, (3) Average, (4) Connoisseur, or (5) Expert. 298 • Age: We asked respondents for their year of birth and calculated their age as of their birthday 299 in 2021. All respondents had to be the legal drinking age in the U.S. (minimum 21) at the time of the 300 survey. 301 • Income: We asked respondents to report their annual household income as (1) under $35,000, 302 (2) $35,000 - $49,999, (3) $50,000 - $74,999, (4) $75,000 - $99,999, (5) $100,000 - $149,000, or (6) 303 $150,000 or more. 304 • Education: We asked respondents to report their highest level of completed education as (1) 305 High school graduate or less, (2) Some college, (3) Completed technical/2-year degree, (4) Completed 306 4-year degree, (5) Some graduate school, (6) Completed graduate degree, or (7) Completed terminal 307 degree. 308 Six respondents chose “Prefer not to answer” for Income and four others were missing another data 309 point and were not included in the regression analysis. We tested the assumptions of regression and 310 there were no issues across the three regressions. We found that collinearity between the independent 311 variables was not an issue, as variance inflation factors (VIF) ranged from 1.097 to 1.366. 312 313 4. RESULTS 314 315 The respondents did not have a strong interest in ingredient and nutritional label information in 316 general. Less than half of the sample said that they read label information Often or Very Often, 317 WEP – Wine Economics and Policy Just Accepted Manuscript 12 whether it was ingredients (Often: 31.4%; Very Often: 17.9%) or nutritional information (Often: 318 31.1%; Very Often: 18.6%). When it comes to using nutritional information to decide which alcoholic 319 beverage to consume, or whether to consume one, barely a quarter (25.3%) said that it was Important 320 or Very Important. The lack of a strong interest in general ingredient and nutritional information 321 carries over to wine even though, overall, the respondents think that wine is associated with good 322 health. When asked which alcoholic beverages, if any, are associated with a healthy lifestyle or diet, 323 almost 75% selected wine. When asked if they would agree that moderate wine consumption is good 324 for health, the mean response was 3.958 on a 5-point scale. 325 326 4.1 Ingredient information when choosing a wine 327 328 Overall, respondents think that knowing the ingredients when deciding which wine to purchase is 329 somewhat important, with a mean response of 3.037. Table 4 presents the regression results for the 330 question “For each of the wine purchase occasions listed, indicate how important it would be to you 331 to know what the ingredients are in deciding which wine to buy” (Ing-Choice). 332 The model was a significant predictor of Ing-Choice (F (8, 293) = 10.652, p < 0.001), accounting for 333 20.4% of the variance in the model. Price, Age, Physical Activity, Diet, Education, and Wine 334 Consumption were all significant predictors of wanting to know ingredients when deciding which 335 wine to purchase. 336 The higher the price category the respondent reported purchasing at least monthly, the more the 337 importance of knowing the ingredients increased (p < 0.001). Those who were more active (p = 0.019) 338 or had a healthy diet (p = 0.014) were more likely to want to know the ingredients in deciding which 339 wine to purchase, and Core wine consumers wanted to know the ingredients more than Marginal wine 340 consumers (p = 0.077). On the other hand, wanting to know the ingredients decreased with age (p = 341 0.004) and education (p = 0.058). 342 343 Table 4. Regression results for dependent variable Ing-Choice 344 B SE t Sig. Wine Consumption 0.223 0.126 1.772 * Price 0.264 0.069 3.840 *** Physical Activity 0.308 0.131 2.356 ** Diet 0.314 0.127 2.483 ** Wine Knowledge 0.167 0.185 0.898 WEP – Wine Economics and Policy Just Accepted Manuscript 13 Age -0.010 0.003 -2.933 *** Income 0.052 0.039 1.328 Education -0.070 0.037 -1.905 * Constant 2.830 0.284 9.974 *** F (8, 293) 10.652 *** Note: *, **, and *** indicate significance levels (two tailed) of 0.10, 0.05, and 0.01. N = 302. Adjusted R2 = 0.204. 345 4.2 Ingredient information when determining willingness to pay 346 347 On average, respondents think that knowing the ingredients when deciding how much to pay for a 348 wine is slightly less important than when deciding which wine to purchase. The mean response for 349 this variable was 3.014. Table 5 presents the regression results for the question “For each of the 350 following wine purchase occasions listed, indicate how important it would be to you to know what 351 ingredients are in a bottle of wine in deciding how much you are willing to pay for the wine” (Ing-352 Pay). 353 The model was a significant predictor of importance of knowing ingredients for willingness to pay 354 (F(8, 293) = 8.046, p < 0.001) and accounted for 15.8% of the variance in the model. Similar to the 355 importance of knowing ingredients when deciding which wine to purchase, Price (p = 0.001), Diet (p 356 = 0.008), and Age (p = 0.017), are significant predictors of wanting to know ingredients when 357 deciding how much to pay for a wine, with the importance of knowing the ingredients increasing with 358 the level for each variable except Age. When deciding how much to pay, Wine Knowledge is also a 359 significant (p = 0.030) positive indicator of wanting to know the wine’s ingredients. 360 361 Table 5. Regression results for dependent variable Ing-Pay 362 B SE t Sig. Wine Consumption 0.117 0.136 0.864 Price 0.248 0.074 3.351 *** Physical Activity 0.226 0.140 1.607 Diet 0.362 0.136 2.654 *** Wine Knowledge 0.435 0.199 2.179 ** WEP – Wine Economics and Policy Just Accepted Manuscript 14 Age -0.009 0.004 -2.398 ** Income 0.015 0.042 0.347 Education -0.051 0.039 -1.289 Constant 2.943 0.305 9.644 *** F (8, 293) 8.046 *** Note: *, **, and *** indicate significance levels (two tailed) of 0.10, 0.05, and 0.01. N = 302. Adjusted R2 = 0.158. 363 4.3 Nutrition information when choosing a wine 364 365 Collectively, respondents were less interested in knowing nutrition information than in knowing a 366 wine’s ingredients. The mean response for the Nutrition variable was 2.481. The regression results 367 for the importance of nutrition information for wine are reported in Table 6, which looks at the 368 questions related to “For each of the following nutritional items, indicate how important you think 369 that item is when considering which wine to purchase.” The model was a significant predictor of the 370 importance of nutrition information (F (8, 293) = 10.175, p < 0.001) accounting for 19.6% of the 371 variance in nutrition information. Price (p = 0.001), Physical Activity (p = 0.088), Diet (p = 0.008), 372 and Wine Knowledge (p = 0.072) were positive predictors while as Age increases the desire for 373 nutrition information decreased (p < 0.001). 374 375 Table 6. Regression results for dependent variable Nutrition 376 B SE t Sig. Wine Consumption 0.136 0.134 1.019 Price 0.237 0.073 3.257 *** Physical Activity 0.236 0.138 1.712 * Diet 0.359 0.134 2.682 *** Wine Knowledge 0.355 0.196 1.808 * Age -0.016 0.004 -4.401 *** Income -0.011 0.041 -0.261 Education -0.050 0.039 -1.298 WEP – Wine Economics and Policy Just Accepted Manuscript 15 (Constant) 2.839 0.300 9.452 *** F (8, 293) 10.175 *** Note: *, **, and *** indicate significance levels (two tailed) of 0.10, 0.05, and 0.01. N = 302. Adjusted R2 = 0.196. 377 378 5. DISCUSSION AND CONCLUSIONS 379 380 5.1 Overall summary of results and implications 381 382 Price is the only variable that was highly significant (p ≤ 0.01) across all three regression models. 383 Consumers who purchased a higher-priced wine at least once a month were more interested in having 384 wine ingredient and nutrition information. This result is consistent with the WMC Communications 385 Study [2] that indicated that high-end wine buyers tend to want more information about wine in 386 general and are more likely to agree that the information found on wine labels rarely helps them 387 choose a wine. This group represents a small portion of wine consumers. In our survey, less than 30% 388 of respondents said they purchase a bottle of wine at least monthly at a price of $15 or more per bottle, 389 and almost half of those did not purchase a bottle priced at $25 or more at least once a month. The 390 U.S. Wine Consumer Segmentation study [17] also found that almost half of regular wine consumers 391 say the never buy a bottle of wine in the $25.00 - $29.99 range and almost three-quarters never pay 392 more than $50.00 a bottle. 393 Age was highly significant in two of the regression models (Ing-Choice and Nutrition) and significant 394 (p ≤ 0.05) in the third (Ing-Pay). It is the most significant factor when considering nutrition 395 information. Younger consumers had more interest in ingredient and nutrition information than older 396 consumers. Younger consumers were also more likely to accept getting this information through 397 technology than the labels on a wine bottle. When presented with the statement in our survey, 398 “including the website (URL) or a QR code that links to that information would be a good alternative 399 to listing the ingredient or nutrition information directly on the bottle,” almost 70% of respondents 400 age 40 or younger chose either Agree or Strongly Agree, a significantly higher rate than respondents 401 between 41 and 64 (55%) and 65 and older (41%) [Χ2(8) = 23.336, p = 0.003]. 402 Diet is highly significant for Ing-Pay and Nutrition and significant for Ing-Choice. Respondents who 403 indicated having a healthy diet were more interested in ingredient and nutrition information than those 404 who do not. Similarly, respondents who say that they were more physically active than their peers 405 WEP – Wine Economics and Policy Just Accepted Manuscript 16 were more interested in ingredient and nutrition information, although that variable was only 406 significant for Ing-Choice and marginally significant (p ≤ 0.10) for Nutrition. This result is consistent 407 with the finding of Bazzani, et al., [5] that health consciousness was positively related to the use of 408 wine labels information and the finding of Grunert, et al., [7] that interest in health is a predictor, but 409 not the strongest one, of information wants for nutrition and ingredients. 410 Similar to Annunziata, et al., [10] we found that better wine knowledge (as self-assessed by 411 respondents) is positively related to wanting more information. However, the variable was only 412 significant for Ing-Pay and marginally significant for Nutrition. It is worth noting that the WMC 413 Communications Study [2] indicated that more knowledgeable wine consumers tend to want more 414 information of all kinds about wine and were less likely to use that information in making a wine-415 buying decision. 416 Surprisingly, frequency of wine consumption was not an important factor. Core wine drinkers were 417 more likely to want ingredient and nutrition information, but the variable only reached marginal 418 significance and only in the Ing-Choice model. In contrast, Escandon-Barbosa and Rialp-Criado [9] 419 found that expert wine consumers, defined by the amount and frequency of wine consumption, make 420 more use of nutrition information than non-experts. 421 The education level of respondents was negatively related to the interest in ingredient and nutrition 422 information but, like wine consumption, only reached marginal significance in the Ing-Choice model. 423 Income was the only variable not to be at least marginally significant in at least one regression model. 424 425 5.2 Policy implications of results 426 427 While some wine industry professionals and wine writers advocate for ingredient and nutrition 428 information disclosure (e.g., Pellechia [1] and Gray [3]), the primary push for government regulations 429 has come from the public health sector. In 2007, the TTB issued “Labeling and Advertising of Wines, 430 Distilled Spirits and Malt Beverages; Proposed Rule” [18] that, if enacted, would have required 431 alcoholic beverages covered by the rule to disclose “on any label affixed to the container” the alcohol 432 by volume and a statement of calories, carbohydrates, fat, and protein. The proposed rule notice noted 433 that almost 4 ½ years earlier the TTB had received a petition calling for such disclosure, and more, 434 from the Center for Science in the Public Interest, the National Consumer League, 67 other 435 organizations, and eight individuals (including four deans of schools of public health). 436 Government regulations should weight the costs of the regulations against the expected benefits. Our 437 study suggests the actual benefits of such regulation may be less than the intended benefits. Overall, 438 the respondents think that wine is associated with good health. Still, respondents had only a marginal 439 WEP – Wine Economics and Policy Just Accepted Manuscript 17 interest in having ingredient and nutrition information for wine. Our findings would suggest that the 440 benefit of requiring ingredient and nutrition information on wine bottle labels seems small, especially 441 given research that shows that people tend to use nutrition labels at lower rates than they claim and 442 that having such information often does not change consumers’ choices. For example, Grunert, et al. 443 [19] demonstrated that self-reported use of nutritional labels may be overstated by 50% and that the 444 lack of use is mostly not attributable to not understanding the information on the label. In addition, 445 their results do not prove that the label information changed consumers’ choices, compared to a 446 situation where such information is not available or is not read by the consumer. Furthermore, Köster 447 [20] showed that many food and beverage purchase and consumption decisions are based on routine, 448 habit, or other subconscious factors. 449 The operational cost for wineries, on the other hand, would be significant given the additional testing 450 and chemical analysis that would be required and the cost of having to create new labels and seek 451 TTB (and in some cases state) label approval with each new vintage as nutritional properties change 452 from year to year. One might expect that ingredient and nutrition labeling could lead to negative news 453 stories based on ignorance and fear rather than science and fact, and these stories could result in costs 454 of lost opportunities, especially considering the findings of Pabst, et al. [6], and current TTB 455 regulations related to advertising health claims could make it difficult for wineries to respond to such 456 stories. 457 As EU regulations come into effect, researchers will have opportunities to study the impact of the 458 regulations in the real-world settings that Robinson, et al., [16] concluded would be necessary to 459 produce studies with high evidential value. In the meantime, our study adds to the body of research 460 that calls into question the efficacy of requiring wine ingredient and disclosure information to meet 461 public health goals and suggests that the TTB could benefit from the opportunity to learn from the 462 EU’s experience before issuing its own regulations. 463 464 5.3 Limitations and Research Opportunities 465 466 This study was based on a survey that asked respondents about their interest in having ingredient and 467 nutritional information available. We did not attempt to measure the extent to which they truly would 468 use ingredient and nutritional information in making wine purchase or consumption decisions or how 469 having ingredient and nutritional information would change such decisions. 470 We approached our study from a public health perspective rather than a marketing perspective. We 471 did not investigate whether consumers would be willing to pay more for wine that discloses ingredient 472 or nutrition information. Likewise, we did not study consumers’ preference for ingredient or nutrition 473 WEP – Wine Economics and Policy Just Accepted Manuscript 18 information if having that information would require them to make a tradeoff between having access 474 to this information or some other information, such as food pairings or a description of the wine, that 475 they may use in making wine purchase and consumption decisions. These are all avenues for future 476 research on this subject. 477 478 479 REFERENCES 480 481 [1] “Ingredient Labeling May Soon Show Up on Your Favorite Wine,” Thomas Pellechia, Forbes, 482 May 22, 2020, https://www.forbes.com/sites/thomaspellechia/2020/05/22/ingredient-labeling-may-483 soon-show-up-on-your-favorite-wine/?sh=69f88aa3654c (accessed 10/22/2021). 484 [2] “2020 Wine Market Council Communications Study,” Wine Market Council, September 25, 485 2020, 486 https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwjr9pGAyt487 7zAhUBCTQIHXqrBXYQFnoECAkQAQ&url=https%3A%2F%2Fwinemarketcouncil.com%2Fw488 p-489 content%2Fuploads%2Fdlm_uploads%2F2020%2F09%2FWMC_Communications_Study_Final_9490 -25-2020.pdf&usg=AOvVaw0Yu88_ZVju71laSmSW0UWE (accessed 10/22/2021). 491 [3] “Time for US Wine to Follow the EU,” W. Blake Gray, wine-searcher.com, May 24, 2020, 492 https://www.wine-searcher.com/m/2020/05/time-for-us-wine-to-follow-the-eu (accessed 493 10/23/2021). 494 [4] “Ingredient Labeling: A Winemaker Strikes Back,” Adam Lee, wine-searcher.com, May 28, 495 2020, https://www.wine-searcher.com/m/2020/05/ingredient-labeling-a-winemaker-strikes-back 496 (accessed 10/23/2021). 497 [5] Bazzani C, Capitello R, Ricci EC, Scarpa R, Begalli D. Nutritional Knowledge and Health 498 Consciousness: Do They Affect Consumer Wine Choices? Evidence from a Survey in Italy. 499 Nutrients. 2020; 12(1):84. https://doi.org/10.3390/nu12010084. 500 [6] Pabst E, Corsi A, Vecchio R, Annunziata A, Loose S. Consumers’ reactions to nutrition and 501 ingredient labelling for wine – A cross-country discrete choice experiment. Appetite. 2021; 156, 502 104843. https://doi.org/10.1016/j.appet.2020.104843. 503 [7] Grunert K, Hieke S, Juhl H. Consumer wants and use of ingredient and nutrition information 504 for alcoholic drinks: A cross-cultural study in six EU countries. Food Quality and Preference. 2018; 505 63:107. https://dx.doi.org/10.1016/j.foodqual.2017.08.005 506 WEP – Wine Economics and Policy Just Accepted Manuscript 19 [8] Annunziata A, Pomarici E, Vecchio R, Mariani A. Do Consumers Want More Nutritional and 507 Health Information on Wine Labels? Insights from the EU and USA. Nutrients. 2016; 8(7):416. 508 https://doi.org/10.3390/nu8070416. 509 [9] Escandon-Barbosa D, Rialp-Criado J. (2019). The impact of the content of the label on the 510 buying intention of a wine consumer. Frontiers in Psychology. 2019; 9:2761. 511 https://doi.org/10.3389/fpsyg.2018.02761 512 [10] Annunziata A, Pomarici E, Vecchio R, Mariani A. Nutritional information and health 513 warnings on wine labels: Exploring consumer interest and preferences. Appetite. 2016; 106: 58. 514 https://doi.org/10.1016/j.appet.2016.02.152. 515 [11] Pabst E, Szolnoki G, Loose S. The effects of mandatory ingredient and nutrition labelling for 516 wine consumers – A qualitative study. Wine Economics and Policy. 2019; 8(1):5. 517 https://doi.org/10.14601/web-8216. 518 [12] Pabst E, Szolnoki G, Loose S. How will mandatory nutrition and ingredient labelling affect 519 the wine industry? A quantitative study of producers’ perspectives. Wine Economics and Policy. 520 2019; 8 (2):103. https://doi.org/10.1016/j.wep.2019.05.002. 521 [13] Hayward L, McSweeney M. Investigating caloric values and consumers’ perceptions of Nova 522 Scotia rosé wines. Food Research International. 2020; 127, 108761. 523 https://doi.org/10.1016/j.foodres.2019.108761 524 [14] Hayward L, Finlay E, Lafortune M, Strother H, Tomchuk A, Selviz V, McSweeney M. 525 Investigating the disclosure of ingredient lists impact on consumers' sensory perceptions of red wines 526 produced in Nova Scotia, Canada. Journal of Sensory Studies. 2020; 35(6), e12608. 527 https://doi.org/10.1111/joss.12608 528 [15] Vecchio R, Annunziata A, Mariani A. Is More Better? Insights on Consumers’ Preferences 529 for Nutritional Information on Wine Labelling. Nutrients. 2018; 10(11):1667. 530 https://doi.org/10.3390/nu10111667. 531 [16] Robinson E, Humphreys G, Jones A. Alcohol, calories, and obesity: A rapid systematic review 532 and meta‐analysis of consumer knowledge, support, and behavioral effects of energy labeling on 533 alcoholic drinks. Obesity Reviews. 2021; 22(6), e13198. https://doi.org/10.1111/obr.13198 534 [17] “2019 Wine Market Council U.S. Wine Consumer Segmentation Slide Handbook,” Wine 535 Market Council, September 5, 2019, https://winemarketcouncil.com/download/914/ (accessed 536 10/24/2021; membership required). 537 [18] “Labeling and Advertising of Wines, Distilled Spirits and Malt Beverages; Proposed Rule.” 538 72 Fed. Reg. 41859 (July 31, 2007). https://www.govinfo.gov/content/pkg/FR-2007-07-31/pdf/E7-539 14774.pdf 540 WEP – Wine Economics and Policy Just Accepted Manuscript 20 [19] Grunert K, Wills J, Fernández-Celemín L. Nutrition knowledge, and use and understanding 541 of nutrition information on food labels among consumers in the UK. Appetite. 2010; 55(2): 177. 542 https://doi.org/10.1016/j.appet.2010.05.045. 543 [20] Köster E. Diversity in the determinants of food choice: A psychological perspective. Food 544 Quality and Preference. 2009; 20(2): 70. https://doi.org/10.1016/j.foodqual.2007.11.002. 545