A Certification for Natural Wine? A Comparative Analysis of Consumer 3 Drivers in Italy and Spain

33 The 2020 certification of natural wine (NW) in France has unleashed a heated debate in Europe. 34 However, knowledge about NW consumer profiles and preferences in a comparative perspective 35 remains scarce in the academic literature. This study aims to define the perceptions, preferences and 36 profiles of wine consumers who support a NW label. For this purpose, we employed analysis of 37 variance, aprioristic factor analysis and multiple regression analysis to examine data from a direct 38 survey performed in Italy and Spain in 2020. Findings reveal that NW consumers in both countries 39 deem it necessary to establish a certification for NW. However, we found significant differences 40 regarding consumers’ profiles, as well as purchasing preferences. In Spain, demand for NW 41 certification is linked to eco-healthy and proximity-craft attributes of wine, and is considered more 42 important by non-professional consumers and those with lower educational level. In Italy, 43 information on the label and the purchase experience are the most important factors to aid in 44 recognizing NW, while women show a significant interest in the NW certification. These findings 45 may help policy-makers to establish homogeneous parameters to differentiate and certify NW. 46 47


In the case of NW, consumers still do not have a clear notion of what it is, how it differs from other 128
wines, and where to buy it [6,14]. In fact, the lack of NW certifications and the uncertainty associated 129 with its attributes have generated a very unstable market for this product in traditionally wine-130 producing countries such as Spain, where consumer research about NW is scarce [15,16]. Recent 131 research has shown growing consumer interest in products with environmental attributes, such as 132 organic, healthy, sustainable, responsible and "proximity" wines [17,18]. It is accompanied by a 133 surge in new certifications that aim to convey trust and inform about wine's intrinsic and extrinsic 134 qualities [19]. Previous research has shown that consumers are willing to pay a premium for organic 135 wines in the belief that they are healthier, tastier, and of higher quality [3,20]. However, the 136 differences between organic, biodynamic, or sulfite-free wines remain confusing for many in the face 137 of the proliferation of information associated with the naturalness of these wines [21][22][23][24]. Fewer 138 studies have been conducted about NW specifically, although the literature on the topic is rapidly 139 growing. Such coverage focuses on the different productive models in the natural wine field [15,25,140 26] and on the policy controversies arising from the recent French certification of natural wine "vin 141 mèthode nature" [27]. Studies focusing on consumer interest in NW show that a predisposition to 142 pay more for NW by Italian consumers was associated with drinking frequency and occasion, organic 143 production, sulphite content, income, and attitudes towards healthy eating and the environment [28, 144 29]. Other studies have explored consumer perception of NW in the Italian wine industry [14], the 145 construction of taste in the NW market [30], and marketing strategies in the NW sector [2,25]. There 146 is a research gap regarding comparative studies between countries, in that we must better understand 147 consumer behavior and specific socioeconomic profiles, given the current dominant focus on Italy in 148 the literature. This study therefore enriches the growing literature on sustainable food consumption 149 attitudes. A further research gap lies in the lack of studies on consumer interest in NW certification, 150 a timely issue with normative implications because the certification developed by the French 151 government has changed the field of play. Owing to these present academic gaps in knowledge, this 152 manuscript sets out to answer two key research questions. First, is there a specific consumer profile 153 that demands the differentiation of NW in the market? Second, what are the preferences and 154 perceptions of wine consumers who deem a certification to identify NW important? Answering these 155 questions is a prerequisite for the development of a EU-wide certification and to better address both 156 producers and consumers' needs and expectations, and help NW producers to create new marketing 157 communication strategies. Owing to the present academic gap in knowledge to date, this paper deals 158 with consumer profiles in relation to NW label interest and their preferences regarding a NW 159 certification. For this purpose, data were collected through a direct survey delivered in Italy and 160 Spain. Both are traditional producing and consuming wine countries leading in terms of vineyard 161 surface area, production volume and export value rankings worldwide, only after France [31]. 162 Ultimately, the paper offers an original contribution to a rather unexplored but emerging topic. As mentioned, a convenience sampling procedure was applied in the absence of a regular wine 180 consumer population census. Eligibility was based on the definition of regular wine drinkers by Wine 181 Intelligence [32], i.e., individuals consuming wine at least once a month. This description has been 182 previously used in similar wine consumer studies [33-35], The questionnaire was launched through 183 specialized sector agents in both Spain and Italy. Producers' associations, distributors, wine critics, 184 sommeliers, wine observatories, etc., publicized the initiative through their websites. They requested 185 the participation of their users, clients and followers, to improve the response rate among wine 186 consumers in both countries. Through this system, a total of 527 fully completed surveys by wine 187 consumers were collected in Spain and 501 in Italy during the two months from mid-September to 188 mid-November 2020. 189

Data analysis 190
Data analysis consisted of two phases in order to understand the drivers influencing demand for NW 191 labeling. Consumer demand for a NW label was our dependent variable and was assessed by asking 192 "To what extent do you consider labeling important to identify NW?". Respondents answered this 193 question using a five-point scale (from 1 "not at all" to 5 "a lot"). The first phase of the analysis 194 responds to our research question regarding profiles of consumers that expressed a need for NW The assumption of normality is used especially when any of the factor categories has less than 50 202 cases. It was tested through Kolmogorov-Smirnov or Shapiro-Wilk statistics and was not fulfilled in 203 several ANOVAs. Therefore, rejection of the hypothesis of equality of means was replicated by 204 default through a Kruskal-Wallis test. For the case of the independent dichotomous variables (gender 205 and NW consumption), means were compared by evaluating the level of significance associated with 206 the F value [37]. 207 The purpose of the second phase was to answer our second research question, about the preferences 208 and perceptions of wine consumers who consider labels important as a means to identify NW. First, 209 it included r Pearson correlations (p <0.01), in order to discriminate between independent and 210 quantitative variables (IV) in wine labeling information, and on purchasing occasions that best 211 correlated with the DV for each country. It also established a ranking order and a comparison between 212 the two. All the 85 IV in the questionnaire were used to prepare this ranking. The aim of this bivariate 213 exploratory statistical analysis is to identify the best IVs that explain the DV in Spain and Italy. This 214 also sheds light on the differences between the two countries. The exploratory bivariate analysis was 215 followed by an aprioristic factor analysis to group the best IVs from each country under common 216 latent dimensions. This strategy permits such exploratory factor analysis and makes it more efficient, 217 thanks to avoiding the rubbish in, rubbish out phenomenon described by [38], which can result from 218 factoring in an indiscriminate number of variables. 219 The factor extraction method is based on principal component analysis using a Varimax rotation. In 220 all cases, Bartlett's test of sphericity rejected the null hypothesis that the observed correlation matrix 221 is an identity matrix (p < 0.01), which legitimizes aprioristic factorizations [38]. The eigenvalues obtained for each of the factors created are always greater than the unit. The scores obtained in 223 differential format for each factor are calculated using the regression estimation method. Finally, the 224 factors are used as IVs in a multiple regression model aimed at explaining the largest percentage of 225 variance in the DV and establishing an explanatory or predictive model for each context. Use of the 226 factors in the explanatory model was supported by a corresponding significant F value (p <0.01) [39]. 227 In the multiple regression models, we checked the assumptions of normality and homoscedasticity of 228 the residuals, as well as the linearity of the IVs with respect to the DV. It was also verified that the 229 Durbin-Watson statistic was between 1.5 and 2.5 in the models, so the residuals were mutually 230 independent. 231 232

Socio-demographic profiles & NW certification 234
This section responds to our question about whether there is a specific consumer profile correlated 235 with demand for the differentiation of NW in the market. Table 1 shows the sample description from 236 both countries. The Spanish sample was composed of 527 wine consumers, 69% men and 31% 237 women, with a mean age of 45 years. More than 50% of the Spanish respondents had a net monthly 238 income of between 1,001 and 2,000 € and university or master studies (79.4%). In addition, 36.2% 239 of the population surveyed considered themselves to be wine professionals. Regarding wine 240 consumption habits, 44.2% of respondents drank wine several times a week, 78.0% consumed NW 241 (33.3% at least once a month), and considered that a certification for NW is necessary (3.95 out of 242

5). 243
The Italian sample was made up of 501 wine consumers, 55% of whom were men and 45% women, 244 with an average age of 38. Almost 50% of the Italian respondents had a net monthly income between 245 1,001 and 2,000 € and university or master studies (55.5%). Finally, 19.6% of respondents defined 246 themselves as wine professionals. Regarding wine consumption habits, 36.7% drank wine several 247 times a week, 68.7% consumed natural wine (32.3% at least once a month) and considered a NW 248 certification necessary (3.90 on average out of 5). 249 250

252
The significant sociodemographic profiles or IV associated with the desire for a NW label in each 253 country are detailed in Table 2. Results show that in both countries NW consumers are more likely 254 to ask for a certification, to be able to identify it in the market. However, ANOVA shows significant 255 differences between consumer profiles from Spain and Italy. 256 257

261
In Spain it was the non-professional consumers and those without university or higher education who 262 most expressed a need for NW certification. In Italy, on the other hand, (female) gender was the 263 variable that discriminated the importance of such certification. All these variables showed significant 264 differences (Sig. F<0.05), but there were also descriptive differences to consider in both samples. For 265 example, Spanish consumers with lower purchasing power most valued the need to certify NW. This 266 difference in means would be significant by the F test, not by KW. In Italy, younger consumers 267 thought it more important to certify NW, whereas in Spain the over-50s showed the greatest interest 268 in this, although not with statistical significance. 269 270

Wine purchasing preferences, perceptions & NW certification 271
This section responds to our question about the preferences and perceptions of wine consumers who 272 deem a NW certification desirable. In order to understand the underlying data structure, Table 3 shows 273 the r Pearson correlations (p <0.01) that best explain the DV for the cases of Spain and Italy, 274 establishing a comparative ranking between them. 275 276

279
In the Spanish case, ranking results show that considering NW healthier than other wines was the 280 main motivation for its consumption among those who would prefer it had a certification. Secondly, 281 the fact that NW wines are organic and sustainable is the next motivation for their consumption. 282 Third, wineries are the preferred place for NW purchase. Subsequently, the winery name or brand 283 when choosing a bottle of wine appears fourth in the ranking. Following this in fifth place, the fact 284 that NW is handmade is a motivation for its consumption and demand for certification. Finally, the 285 region and country of origin is in sixth position, while wine shops as the preferred place of NW 286 purchase follow this as seventh. This means that for Spaniards who consider a NW certification 287 necessary, it is important to characterize attributes such as healthy, sustainable, ecological and 288 artisanal, which are the main motivations for its consumption. In comparison, these attributes rank 289 20 th , 13 th and 30 th in Italy). 290 In the Italian case, those wishing for a NW certification prioritize the place of purchase over the 291 specific attributes of NW. In other words, in first place they prefer to go to tastings to identify and 292 buy it. In second place, Italians prefer to go to wine shops (seventh in Spain) and, in third place, they 293 rely on books, guides or specialized magazines to identify NW. Biodynamic certification (e.g., 294 Demeter) is important for Italians, appearing fourth in the ranking. It is noteworthy that for Italians 295 who would prefer NW certification, the internet and social networks are important spaces for These variables were subsequently organized through an aprioristic factor analysis to identify clusters 303 of explanatory variables of the DV. These factors have a higher Pearson's r than the IV variables that 304 comprise them (see Table 4), so their predictive capacity will be greater for the DV.

310
In the Spanish case, the first explanatory factor associated with demand for a NW certificate clusters 311 the variables expressing NW attributes that characterize it as healthier, more sustainable and 312 ecological than other wines. This factor was defined as eco-healthy (F1). A second explanatory factor 313 emerges for the Spanish context that combines the perception of NW as artisanal and the preference 314 to buy directly from the winery. We named this factor proximity-craft (F2). A third factor combines 315 the importance of the brand or product name with the region and country of origin when a person 316 supporting NW certification chooses a wine. We called this factor origin-brand (F3). In the Italian 317 case, a factor identified as wine experience (F4) groups together tasting and wine shops as means to 318 identify and buy NW. Another factor defined as on-label-info (F5) groups the importance of organic 319 and biodynamic certifications with brand information, in order to recognize NW. Finally, we defined 320 a sixth factor extra-label-info (F6) as the importance of information widely retrieved in the media to 321 recognize NW, including books, social networks, specialized magazines and similar outlets. 322 Finally, these factors were used as IV in a multiple regression analysis in order to establish an 323 explanatory or predictor model for each country. Table 5 shows the factors that explain a higher 324 percentage of variance for both cases in a combined rather than independent form. This provides 325 robustness to these combinations when explaining the DV (see Table 5 and estimated coefficients in 326 Table A1). 327 328  F change<0.01), that is, the perception that NW is both eco-healthy and 332 proximity-craft. In contrast, the factor associated with origin-brand (F3) was left out of the model as

Discussion 342
Our results in Spain and Italy show that NW consumers are more likely to demand a certification that 343 identifies NW in the market. However, there are differences between the socio-demographic profiles 344 of consumers and the drivers of NW consumption between countries. In Spain, the socio-345 demographic profile of consumers who support NW certification includes non-professional 346 consumers and people with no university or higher education, whereas in Italy, as a group women do. 347 These results converge with recent research by [40], which found that women pay more attention to 348 wine labeling and are more likely to pay for NW in the Italian market. In the Spanish case, there is 349 no previous research on consumer profiles and NW, so further work is required in this area. Traditionally, in many mostly southern parts of Europe, wine was an everyday food staple [45]. 360 However, with the globalization of wine and the proliferation of brands and quality labels, the choice 361 to purchase wine has made the debate about which consumers choose which wine more complex. debating whether to charge wine companies producing over 25,000 bottles per year more, to prevent 412 appropriation of the label by industry. Therefore, aspects related to proximity-craft should also be 413 taken into account when certifying NW. 414 Finally, there is the experiential factor (F4) among those wanting a NW certification in Italy. In 415 general terms, wine has been considered as a unique product and different from any other food whose 416 singularity makes it an experiential product [58]. Wine is ceasing to be a traditional food in rural 417 societies and is becoming more and more a product associated with hedonic or luxury consumption, 418 especially in non-wine producing regions [36]. The recent review on consumer behavior by Deroover 419 et al. [59] highlights that wine is perceived as an expression of traditions and culture. These attributes 420 have greater influence on purchasing and consumption choices than for any other food or beverage 421 [60]. Our results show that the lived-experience of identifying and buying a NW in specialized wine shops and wine tastings is also part of this consumption pattern. Therefore, a NW certificate should 423 incorporate aspects that differentiate these wines with regard to the unique experience that can take 424 place through NW consumption. 425 426

Conclusion 427
This study has furthered demographic and sociocultural knowledge of the consumer profiles and 428 drivers of demand for a NW certification, distinguishing as an example between Spanish and Italian 429 wine drinkers. To answer the initial research question as to whether there is a specific consumer 430 profile that considers a NW label necessary, results show that those who already consume NW are 431 the most likely to demand a certificate to differentiate NW in the market, both in Italy and Spain. 432 However, there are significant differences between these profiles. In Spain, non-professional 433 consumers and those with lower educational levels support a NW certificate to a greater extent, while 434 in Italy it is women who show a greater interest. NW to prefer a distinctive label and that the explanatory factors for such a certificate are not mutually 441 exclusive. Therefore, these results suggest that the stakeholders could indeed set homogeneous 442 standards to reduce uncertainty and information asymmetries concerning NW. This constitutes an 443 important contribution to the debate on the need to establish a common regulatory framework leading 444 to a consensual EU-wide creation of a distinctive NW label. This would help assuage the polemics 445 among countries, as reflected in complaints by several Members of the European Parliament against 446 the French certification initiative Vin Méthode Nature. It could also satisfy NW producers who 447 demand some sort of differentiation in the wine market given the lack of ingredient labeling in wine. 448 In parallel, it could also respond to consumer interests in more transparent and sustainable food 449 products, given that the new regulations for labeling alcoholic beverages are still being debated, with 450 no clear outcome [9]. Certainly, a rise in the minimum standard requirements for ingredient labeling 451 in the wine sector would make NW producers less interested in having their own certification [61], 452 but this scenario is still unclear. Given that the promotion and growth of NW would be positive in 453 terms of environmental sustainability and a cleaner food chain in Europe and beyond, a NW 454 certification is in line with new EU-wide strategies such as From Farm to Fork and the European 455 Green Deal. A pro-labeling policy for NW would be therefore fundamental to reconcile sustainability 456 aims with consumer and producer interests alike, thus reducing the currently prevailing information 457 asymmetry in the wine market. 458 We are aware that the lack of a probabilistic sample is a core shortcoming of this study, which limits 459 the potential to extrapolate the results to the whole wine consumer population in Spain and Italy at 460 large. Furthermore, the explained variance in the regression models is low, which implies that there 461 may be more drivers influencing demand for a NW certificate. Future research should explore the 462 perspectives of other actors in the wine sector on NW labeling, such as winemakers or traders, to 463 further optimize NW communication and marketing. Similarly, more cross-national and comparative 464 studies are needed to better delve into the profile of the NW consumers and the perceptions associated 465 with them, in order to more effectively deliver a Europe-wide certification.