Bio -based and A ppl ied Economics BAE Bio-based and Applied Economics 10(3): 207-218, 2021 | e-ISSN 2280-6172 | DOI: 10.36253/bae-10044 Copyright: © 2021 G. Gastaldello, D. Mozzato, L. Rossetto. Open access, article published by Firenze University Press under CC-BY-4.0 License. Firenze University Press | www.fupress.com/bae Citation: G. Gastaldello, D. Mozzato, L. Rossetto (2021). Drinking Covid-19 away: wine consumption during the first lockdown in Italy. Bio-based and Applied Economics 10(3): 207-218. doi: 10.36253/bae-10044 Received: November 23, 2020 Accepted: July 13, 2021 Published: January 11, 2022 Data Availability Statement: All rel- evant data are within the paper and its Supporting Information files. Competing Interests: The Author(s) declare(s) no conflict of interest. Editor: Davide Menozzi. ORCID GG: 0000-0002-7337-9897 DM: 0000-0002-4130-7290 LR: 0000-0001-7070-153X Authors contributions: All the authors cooperated in producing the manu- script. In particular: D. Mozzato pro- vided guidance in data analysis, while L. Rossetto supervised the research project. L. Rossetto and G. Gastal- dello equally contributed to writing the theoretical background. L. Rossetto, G. Gastaldello and D. Mozzato equally contributed to reviewing the original manuscript. Finally, G. Gastaldello per- formed the data analysis and wrote the remaining paragraphs of the manu- script. Drinking Covid-19 away: wine consumption during the first lockdown in Italy Giulia Gastaldello*, Daniele Mozzato, Luca Rossetto Department of Land, Environment, Agriculture and Forestry (TESAF), University of Padova, Italy *Corresponding author. E-mail: giulia.gastaldello.1@phd.unipd.it; daniele.mozzato@ unipd.it; luca.rossetto@unipd.it Abstract. In Italy, wine is an integral part of most people’s habits and lifestyles. The advent of a traumatic event like the Covid pandemic brought profound changes to peo- ple’s lives: economic instability and normality disruption led consumers to revise their priorities and modify their consumption and purchase behavior. This study analyses the impact of socio-demographic, psychological, and context-related modifications induced by the pandemic on wine consumption and purchase patterns. Participants completed an online, structured survey, and the sample is constituted by Italian wine consumers. Logis- tic regression and descriptive techniques are applied to analyze data. Results highlight that wine consumption is a deeply rooted habit among Italian consumers, which resisted the great context modifications that occurred with the pandemic. Moreover, changes in wine consumption are connected to that of other alcoholic beverages. Significant short- term and potential long-term effects are discussed. Information collected is paramount to understanding wine consumers’ reactions and behavioral changes induced by the pan- demic and effectively plan marketing strategies during new infection peaks. Keywords: Covid-19, wine consumption, Italy, consumer behavior, logistic regression. JEL codes: D12, L66. 1. INTRODUCTION The Covid pandemic is a traumatic event that led to significant changes in people’s lives and Italy is among the hardest-hit European countries (Min- istero Della Salute, 2021). The rapid and insidious spread of the virus, caus- ing a severe and potentially life-threatening respiratory disease, forced the national Government to take a drastic step by forcing the country into a first, extended lockdown that lasted from March 10th to May 4th, 2020. This peri- od had a profound impact on two significant aspects of the national com- munity: on economy and productivity due to the forced shutdown of most activities, and on mobility and social occasions through the prohibition of physical gatherings and trips both outside and within the region. The disrup- tion of people’s habits and lifestyles generated severe psychological discom- forts (Colbert, Wilkinson, Thornton, and Richmond, 2020, Arpaci, Karataş and Baloğlu, 2020). http://creativecommons.org/licenses/by/4.0/legalcode 208 Bio-based and Applied Economics 10(3): 207-218, 2021 | e-ISSN 2280-6172 | DOI: 10.36253/bae-10044 Giulia Gastaldello, Daniele Mozzato, Luca Rossetto The food-and-beverage industry and retail trade were among the few activities allowed to operate by the nation- al law, so access to wine and other alcoholic beverages was still available. Although wine is an essential compo- nent of Italian culture and lifestyle (Seghieri, Torrisi, and Casini, 2007), the lockdown profoundly transformed wine consumers’ routines, leading to potential modifications in wine consumption patterns. Such changes can potentially affect future wine demand, thus the role of wine in Ital- ian culture and lifestyle. Considering the unprecedented circumstances of uncertainty the wine industry is facing, there is a need for reliable information on the impact of the lockdown on wine consumption. This study aims at responding to this need, identify- ing factors that triggered modifications of wine consum- ers’ behavior during the first national lockdown. The effect of new consumption habits such as online purchas- ing is also explored to provide insights into whether such factors can affect demand in the long term. An online survey was conducted on a large sample of Italian con- sumers of wine and alcoholic beverages (beer and spirits) to achieve these goals. A descriptive analysis is under- taken to highlight significant changes in alcoholic bever- age consumption during the lockdown, focusing on both wine and substitution effects among wine, beer and spir- its. Finally, factors inducing positive and negative modifi- cations of wine consumption frequency are identified. 2. THEORETICAL BACKGROUND The Covid pandemic is a one-of-a-kind, extraor- dinary event. Although world economies have already experienced health emergencies due to virus outbreaks such as SARS, the Covid pandemic crisis is unprecedent- ed due to the multi-level and interdependent changes it has induced on a global scale. Consequently, to the best of the authors’ knowledge, no similar phenome- na in nature and magnitude have been analysed in the existing literature on consumer behavior. Nevertheless, researchers extensively explored the effect of habit dis- ruption and stress on the consumption of both wine and alcoholic beverages. The following sections present the state of the art on these two key aspects, which are iden- tifiable as significant consequences of the pandemic on consumers’ lifestyles. Additionally, the role of wine and wine consumption in Italy is discussed. 2.1 Wine and alcohol consumption in Italy Italy is among the major players in the wine market, with a production exceeding 50 million hl/year (OIV, 2020), and being the third world wine consumer after the USA and France. The Italian population consumes over 20 million hectoliters a year (22.4 in 2018), corresponding to 35 liters per capita (OIV, 2020). The latter has consider- ably shrunk in the last decades (Sellers and Alampi-Sot- tini, 2016). Still, its decrease is primarily due to changes in the way of consuming wine rather than to a switch of consumer preferences towards other alcoholic beverages. The function of wine has indeed gradually switched from nutrition to pleasure (Hertzberg and Malorgio, 2008), leading the share of daily wine consumers to decrease (17.6% of the population) in favor of non-daily ones, which are growing (36.6% of the population) (ISTAT, 2020). Generally, older generations tend to drink more wine and more often than younger ones (ISTAT, 2020), and contemporary daily wine consumers are likely to be males rather than females (25.9%; ISTAT, 2020). Females, indeed, are generally less prone to alcohol drinking. Com- pared to other alcoholic beverages such as beer and spir- its, wine consumption is to a greater extent rooted in the Italian population’s habits: indeed, only 5.3% and 0.6% of consumers, respectively, are daily beer and spirits drink- ers (ISTAT, 2020). The widely diffused habit of having wine during meals could be an explanation, and makes food and gatherings with family members important con- sumption motivations in the Italian scenario. Social rela- tions also represent a relevant consumption motivation, mainly due to the habit of the pre-meal aperitivo, when wine is consumed either by the glass or mixed in cock- tails. Almost 40% of the Italian population usually drink alcoholic aperitivo (ISTAT, 2020). Finally, health has been thoroughly explored as a factor inf luencing alcohol drinking behavior. Gener- ally, alcohol consumption can be either considered a potential threat because of the poisonous effect of alco- hol overconsumption and the related health risks, or an unhealthy dietary choice, increasing the caloric intake while providing low nutritional value (Bazzani et al., 2020). In this respect, the new post-pandemic lifestyle could prompt a recalibration of priorities (Sigala, 2020), leading people to re-evaluate the outcomes of their behaviors (Wood, Tam, and Witt, 2005). Therefore, the pandemic may constitute a deterrent to alcohol con- sumption, as human health and survival are at stake. As regards to wine past research highlights the potential beneficial effects of moderate wine consump- tion on human health, mostly related to antioxidants in red-colored berries (e.g., Nijveldt et al., 2001; De Lorimier, 2000). The Mediterranean diet can potentially improve such effects. In this sense, wine can be seen as a healthy dietary choice (Fiore, Alaimo, and Chkhar- tishvil, 2019). 209Drinking Covid-19 away: wine consumption during the first lockdown in Italy Bio-based and Applied Economics 10(3): 207-218, 2021 | e-ISSN 2280-6172 | DOI: 10.36253/bae-10044 On a broader scale, a moderate wine intake can also contribute to hedonistic health and well-being (Fiore, M., Alaimo, L. S., & Chkhartishvil, N., 2019), which is associated with focusing on the self and the present moment (Huta, 2015), favoring sociability and inducing a stress-free mood (Cooper, 1994). The unprecedent cir- cumstances of the lockdown and the uncertainty gener- ated by the pandemic may have emphasized the role of wine drinking in emotional and mental well-being, posi- tively impacting on its consumption frequency. 2.2 Psychological difficulties and alcohol consumption behavior As mentioned above, the Covid-19 pandemic con- stituted a source of stress and anxiety that have long been associated with increased alcohol consumption. Among the first theories, there is Horton’s Tension- reduction hypothesis (1943), identifying alcohol con- sumption as a way to diminish the feeling of anxiety prompted by stress, which arises either from traumatic events or from environmental stressors. Later studies further explored this connection, specifying that alco- holic beverages consumption can be a way to mitigate negative feelings (Powers and Kutash, 1985). Stress, moreover, may increase alcohol intake when the inten- tion to drink is already present (Dawson, Grant, Stin- son, and Zhou, 2005). In this respect, a stronger asso- ciation exists with being male (Dawson et al., 2005), which can be explained by gender-related stress resist- ance. Indeed, men and women tend to react differently both to stress and to single stressors, intended as fac- tors inducing stress (APA, 2012). As regards age, older people tend to deal better with negative emotions when subject to stressors, which is believed to be an indica- tor of stress resilience (Ong et al., 2006), while mainly endorsing positive feelings (Scott et al., 2014). Never- theless, it should be considered that aging is associ- ated with increased emotional complexity connected to the awareness of “running of time” (Carstensen et al., 2000). Such complexity peaks at middle age (Labouvie- Vief et al., 2007). In the context of this research, we expect the fear of the SARS-CoV-2 illness, jointly with economic uncer- tainty and isolation, to trigger an increase in wine and, more generally, in alcohol consumption frequency, espe- cially in males and middle-aged people. Indeed, other researchers have already outlined the high risk of poten- tial alcohol overconsumption prompted by the pandem- ic’s emergence (e.g., Clay and Parker, 2020). 2.3 Disrupting (wine) habits Habits are defined as behavioral dispositions to repeat a set of everyday activities when specific circum- stances occur (Wood et al., 2005). As an individual repeats the behavior, triggering factors – e.g., perfor- mance time, location, or people the activity is usually shared with – are associated in the memory with specific activities, leading to a set of cognitive, neurological, and motivational changes. Habits, indeed, tend to be context- dependent (Wood et al., 2005). Consequently, habitual actions lose their explicit instrumental nature, separat- ing them from intentions (Neal, Wood, and Quinn, 2006; Wood et al., 2005) and being performed almost uncon- sciously. As regards wine, its recurrent consumption in Italy is strongly connected with a multitude of attitudes, behaviors, and consumption situations (Presenza, Min- guzzi, and Petrillo, 2010), mainly conviviality, e.g., the aperitivo or gatherings with friends, colleagues, family and dining out. In this sense, wine drinking can be con- sidered a habit for a large slice of the population. With the closure of restaurants, cafes and the ban of social gatherings during the lockdown, most factors driving wine consumption habits disappeared inducing changes in most wine consumers’ drinking habits. Mainly, peo- ple usually drinking wine on social occasions may reduce their consumption frequency. At the same time, consum- ers who were used to consume wine alone are expected either to keep their consumption frequency stable or to increase it. The direction of this change may differ based on the strength of the role of wine in one’s habits. The literature highlights that stronger habits survive context changes as intentions may come into play, creating the conditions to preserve them (Wood et al., 2005). Given the strength of the habit of aperitivo in Italy, people may look for alternative ways to pursue this activity during the lockdown: the virtual aperitivo. To sum up, the strength of wine consumption as a habit among the Italian population before the pandemic leads to assuming that intentions arose to preserve this, despite the drastic context changes. In this respect, phys- ical barriers imposed with the national lockdown may prompt the emergence of new ways to maintain usual wine consumption habits. Conversely, the lockdown may show disrupting effects on wine drinking, resulting either in a reduced consumption or in substitution effects. 3. MATERIALS AND METHODS A structured questionnaire was developed focusing primarily on wine while incorporating information on 210 Bio-based and Applied Economics 10(3): 207-218, 2021 | e-ISSN 2280-6172 | DOI: 10.36253/bae-10044 Giulia Gastaldello, Daniele Mozzato, Luca Rossetto the consumption of other alcoholic beverages. The sur- vey includes seven sections: consumption and purchase patterns before and after the pandemic (for wine, beer, and spirits), wine consumption context pre- and post- Covid, online wine-related interactions, psychological difficulties (i.e., feeling of isolation, fear of the virus and the economic crisis), positive feelings (i.e., willingness to support local wine producers, possibility to refocus on the self while in lockdown) and socio-demographics. Specifically, isolation is expressed as a latent construct focusing on relational connectedness. Indeed, relational connectedness represents social loneliness, one of the most significant consequences of the lockdown. For this purpose, a 3-items scale based on Hawkley et al.’s (2005) loneliness scale (UCLA scale) was adopted. The 3 items were reduced to a single factor (α=.87; KMO=.72) and the resulting Isolation scale is inverted (1=strong isola- tion; 5= weak isolation). The three items used are “since the beginning of the lockdown, there are people I feel close to”, “since the beginning of the lockdown, there are people I can talk to”, and “since the beginning of the lockdown, there are people I can turn to”. Fear of Cov- id-19 was captured by the statement “I feel vulnerable to Covid-19 outbreak”, while fear of the economic crisis is captured by the statement “I am concerned about the economic impacts of the crisis on myself and my fam- ily”. Two statements represented positive feelings: “quar- antine has allowed me to focus on the essentials”, and “since the quarantine has begun, I feel like I should buy more local wine to support my country’s economy”. All items were measured by 7-points likert scales (1= strong- ly disagree; 7=strongly agree). Online data collection was carried out between April 16th and April 29th, 2020 – i.e., during the first lockdown in Italy. As previously mentioned, given the impossibility of reaching the population of interest – consumers of alcoholic beverages – due to the ongoing pandemic and the short time window available, snowball sampling was adopted. This technique represents an effi- cient and cost-effective data collection method in con- texts where subjects of interest are challenging to reach (Ghaljaie, Naderifar and Goli, 2017). Data were collected according to the guidelines provided by the Declaration of Helsinki (WMA General Assembly, 2013). Drawbacks reported in the literature connected to this sampling technique, primarily due to its conveni- ence nature, can be compensated by the large sample size. The survey was diffused through social networks and via direct contacts. The original study, designed in collabo- ration with the EuAWE (European Associations of Wine Economists) research group, involved several big play- ers in the wine sector – i.e., Spain, Italy, Portugal and France. The current analysis refers exclusively to the Ital- ian sample, with a total of 1076 valid questionnaires col- lected. Table 1 summarizes the descriptive statistics of the sample. The majority of interviewees (57.8%) are males employed in the service sector (57.4%) and with either good (50.1%) or sufficient (36.2%) economic situation. Almost half of the sample live in an urban context, while 30% come from suburban residential areas. A minor share of respondents live alone, with an average household size of three adults (45.7%) and no children (68.4%). Almost all age groups are homogeneously represented, with a slight predominance of 41-50 year old subjects. The age class of over 70s was poorly represented and was aggre- gated into the 60-70 age group. Similarly, for wine con- sumption frequency before the lockdown (WCONS_B), respondents drinking wine once a month or less were aggregated into one category of occasional consumers. Data analysis relies on descriptive techniques and binary logistic regression (LR), given the categorical nature of the dependent variables (DV) and the use of human- sourced data. This statistical approach was chosen due to its capacity to provide higher robustness when multivari- ate normality assumptions and equal variance-covariance matrices across groups are not met, as commonly happens in social science research (Hair et al., 2019). Specifically, two LR models were developed to identify factors trigger- ing positive (model B), and negative (model A) changes in wine consumption frequency during the first lockdown. For the sake of the analysis, consumers of alcoholic bev- erages who do not drink wine and missing income values were excluded through listwise deletion, thus reducing the sample to 1018 respondents. The enter method was pre- ferred to the stepwise procedure, as the latter tends to pro- duce sample-specific results (Hair et al., 2019). Regressors were selected based on the literature. Variance Inflation Factor (VIF) and Tolerance were used to check for multi- collinearity, and all values were within the recommend- ed thresholds (VIF<5; Tolerance>0.2; Hair et al., 2019). Although the primary aim of the analysis is explanatory, additional fitting diagnostics were performed. Overall predictive accuracy of the models was assessed through Receiver Operating Curves (ROC) and Area Under Curve (AUC). According to Hosmer et al. (2013) thresholds, both models show excellent discrimination power (AUC model A = .82; AUC model B = .87). 4. RESULTS 4.1 Wine consumption: the pre-lockdown scenario Before the beginning of the pandemic, most respondents were regular wine consumers drink- 211Drinking Covid-19 away: wine consumption during the first lockdown in Italy Bio-based and Applied Economics 10(3): 207-218, 2021 | e-ISSN 2280-6172 | DOI: 10.36253/bae-10044 Table 1. Descriptive statistics of the sample.   Frequency %     Frequency % Gender     n° adults (respondent included) Male 622 57.8 1 22 2.0 Female 454 42.2 2 145 13.5 Age 3 492 45.7 18-29 223 20.7 4 199 18.5 30-40 176 16.4 5 161 15.0 41-50 269 25.0 ≥6 57 5.3 51-60 244 22.7 Income   >60 164 15.2 Very problematic 13 1.2 Employment (n=1075) Problematic 76 7.1 Agriculture 146 13.6 Sufficient 390 36.2 Industry 104 9.7 Good 539 50.1 Service 618 57.4 No answer 58 5.4 Student 125 11.6 Motivations for wine consumption Retired 58 5.4 Socialization (yes) 372 34.6 Unemployed 24 2.2 Relax (yes) 290 27.0 Has children Health (yes) 118 11.0 No 736 68.4 Food (yes) 738 68.6 Yes 340 14.6 Taste (Yes) 788 73.2 Res. Area Average bottle price wine before the lockdown (n=1052) Rural 240 22.3 <5€ 148 13.8 Residential 331 30.8 5-10€ 439 40.8 Urban 505 46.9 11-20€ 359 33.4 Freq. digital gatherings during the lockdown 21-30€ 75 7.0 Did not do it 161 15.0 >30€ 31 2.9 Rarely 517 48.0 Has a wine app At least once a week 181 16.8 No 750 69.7 Every day 217 20.2   Yes 326 30.3 Wine consumption frequency before the lockdown Wine consumption frequency in lockdown Never (0) 19 1.8 Less frequent 478 44.4 <1 a month (1) 50 4.6 Unchanged 251 23.3 At least 1 a month (2) 134 12.5 More frequent 347 32.2 At least 1 a week (3) 568 52.8   Daily (4) 305 28.3   Beer consumption frequency before the lockdown Beer consumption frequency in lockdown Never (0) 127 11.8 Less freq. 563 52.3 <1 a month (1) 158 14.7 Unchanged 333 30.9 At least 1 a month (2) 265 24.6 More freq. 180 16.7 At least 1 a week (3) 486 45.2   Daily (4) 40 3.7   Spirits consumption frequency before the lockdown Spirits consumption frequency in lockdown Never (0) 279 25.9 Less freq. 565 52.5 <1 a month (1) 309 28.7 Unchanged 414 38.5 At least 1 a month (2) 260 24.2 More freq. 97 9.0 At least 1 a week (3) 214 19.9   Daily (4) 14 1.3   212 Bio-based and Applied Economics 10(3): 207-218, 2021 | e-ISSN 2280-6172 | DOI: 10.36253/bae-10044 Giulia Gastaldello, Daniele Mozzato, Luca Rossetto ing wine at least once a week (52.8%) or daily (28.3%) (Table 1). Concerning other alcoholic beverages, the majority used to consume beer weekly (48.9%) and spirits sporadically (54.6% less than once a month). Based on ISTAT data, the share of daily wine consum- ers in the sample is higher than the Italian national average (17.6%). Wine was mostly consumed for its taste, paired with food during meals, and to socialize (Table 1). Coherently, it was prevalently drunk with friends and relatives. Adapting the Rabobank wine classification (Heijbroek, 2003), two main segments of wine consumers can be identified based on the aver- age price per bottle: premium wine consumers, pur- chasing wine ranging between 5 and 10 Euros per bot- tle (40.8%) and super-premium ones, who usually buy wines priced between 11 and 20 Euros (33.4%). Accord- ingly, half of the respondents declared a good economic situation (50.1%). Results suggest that the sample com- prises wealthier, higher-end consumers compared to the average Italian population, since market data on domestic wine sales report an average price-per-liter of 3.27 € (IRI, 2009). However, such average price is likely underestimated being referred to off-trade sales in supermarkets and discounts. Such sales channels are usually offering wines at a lower average price point compared to restaurants and enoteche, i.e., Italian spe- cialized wine shops, which are excluded. Accordingly, Enoteche are the third most important shales chan- nel for wine in the sample (45.5%) after cellar door sales (direct sales, 48.0%) and supermarkets (51.2%). Other channels such as e-commerce (12.8%) play a minor role, despite the digitalization level of respond- ents: indeed, 40% declared that they have a wine app on their smartphone. Anyhow, the share of online buy- ers in the sample more than doubles the average data reported in the sector literature for online wine sales in developed countries, which is approximately 5% (Higgins et al., 2015). Additionally, 70% of respond- ents declared a great willingness to support local wine producers in response to the Covid-crisis by preferring domestic wines. 4.2 Psychological difficulties during the lockdown As expected, psychological difficulties are strongly felt by all respondents. In particular, the greatest worry concerns the negative economic impact of the pandemic (79.2%; Table 1). Economic concerns are strong enough to overcome fear of the virus, which, notwithstand- ing its life-threatening nature, is suffered by less than a half of the sample. A feeling of isolation, intended as the impossibility to relate with, talk with, and rely on others, also emerges as a dominant feeling for most respond- ents (69.1%; Table 1), despite most of them not living alone. Positive feelings also emerged, as a large share of interviewees see the lockdown as a chance to refocus on themselves (58.0%). Therefore, among the multitude of negative feelings, respondents could see the bright side of the situation. 4.3 Changes in wine consumption during the lockdown Most of the wine consumers interviewed kept pur- chasing wine during the lockdown (75.7%) without changing the average bottle price (Table 2). However, part of the sample either reduced the average bottle expenditure (34.1%) or completely stopped purchasing the product (22.5%).   Strongly disagree Disagree Neutral Agree Strongly agree Willingness to support Italian wine producers buying national wine 2.6 2.6 24.6 37.7 32.4 Feeling of isolation (inverted) 0.9 3.6 26.4 47.8 21.3 Fear of the economic crisis 0.8 5.8 14.2 46.0 33.2 Fear of the virus 3.8 15.0 33.0 40.1 8.1 Refocus on me during the lockdown 2.4 8.7 30.9 44.3 13.7 Notes: where not specified, n=1076. The % column reports the valid percentage. Table 2. Changes in wine purchase pattern following the lockdown.   Frequency % Wine purchase behavior in lockdown     Does not buy wine 19 1.8 Stopped buying wine 242 22.5 Kept buying wine 815 75.7 Average bottle price variation in lockdown Reduced 367 34.1 Unchanged 647 60.1 Increased 62 5.8 Notes: n=1076. 213Drinking Covid-19 away: wine consumption during the first lockdown in Italy Bio-based and Applied Economics 10(3): 207-218, 2021 | e-ISSN 2280-6172 | DOI: 10.36253/bae-10044 In this regard, the presence of over 22% of respond- ents drinking wine from their personal stock while in lockdown (a 134% increase compared to the pre-pan- demic scenario) suggests that a stop in wine purchases does not necessarily correspond to a reduction in its consumption. Crosstabulations support this hypoth- esis, as the stop in wine purchase is not significantly related to reducing wine drinking frequency during the lockdown (chi-square: .67, p = .418). Among respond- ents who kept purchasing wine, results highlight several changes in their wine consumption habits during the lockdown. First, mobility restrictions and confinement impacted both consumption occasions and buying chan- nels. As can be observed in Table 3, during the lock- down respondents consumed wine mostly with their rel- atives (78.1%) or alone (26.4%). Virtual meetings became an alternative social drinking occasion for 13.5% of the sample. Although the share of respondents is limited, this finding suggests virtual gatherings constituted a tool to keep the social dimension of wine alive since the majority of respondents who drunk wine on such occa- sions also reported socializing as a motivation for drink- ing it (57.2%; chi-square: 38.07, p<0.001). As regards sales channels, the mandatory closure of several business activities inevitably impacted on wine purchase patterns, especially for the large share of con- sumers who used to purchase wine directly from the pro- ducer (48.0%) or specialized wine shops (45.4%) (Table 3). Online wine shoppers increased by 43% (Table 3), 43.4% of whom were first-timers (Table 4). This picture leads us to assume that online wine sales partially counterbal- anced the inaccessibility of most sales channels. The lockdown imposed by the Covid-19 pandemic significantly affected wine consumption frequency as well: 23% of the sample kept drinking wine as often as before the pandemic, while the great majority either increased (32.2%) or decreased it (44.4%). Model A and model B investigate factors impacting the decrease (DV1) and increase (DV2) in wine consumption frequency among wine consumers during the first lockdown (Table 5). First, results reveal that none of the psychological difficulties directly affects neither DV1 nor DV2. Both decrease and increase in wine consumption frequency are connected to a parallel modification of beer (VARBC) and spirits (VARSC), suggesting variations in wine con- sumption are attributable to a change in the overall alco- holic beverages’ consumption pattern. Nevertheless, the effect of beer consumption frequency is considerably greater than that of spirits. Accordingly, when consider- ing the total expenditure for all alcoholic beverages in the lockdown, only that of wine (for model A and B) and beer (for model B) significantly affect the DV. Focusing on model A (DV1), a reduced wine con- sumption frequency is related to a decreased beer consumption (VARBC_RED). Accordingly, these respondents did not increase total expenditure on wine (INCREXP_W) and no significant effects emerge for var- iations in the total expenditure on beer (INCREXP_B). None of the sales channels show significant impact on the DV1. Families with children (CHILDY) are less like- ly to have reduced wine consumption in lockdown. On the contrary, a significant positive effect emerge with age, suggesting that older subjects have greater odds of shrinking their wine consumption frequency in lock- down. Among the reasons for drinking wine, health and relaxation emerge as significant factors impacting DV1: while drinking to relax (R_RELAX) decreases the odds of reducing wine consumption frequency in lockdown, consuming wine for its health properties (R_HEALTH) seems to promote this behavioral modification. Despite Table 3. Evolution of consumption contex and sale channel before and during the lockdown.     BEFORE the LOCKDOWN DURING the LOCKDOWN Δ n. % n. % Wine consumption context         Friends 857 79.6 83 7.7 -90% Family 754 70.1 840 78.1 11% Colleagues 215 20.0 31 2.9 -86% Alone 193 17.9 284 26.4 47% Digital gatherings 11 1.0 145 13.5 1218% Source of the wine consumed     Supermarket 551 51.2 510 47.4 -7% Direct sales 517 48.0 154 14.3 -70% Specialized Wine shop 489 45.4 112 10.4 -77% Online 138 12.8 198 15.8 43% Personal wine stock 103 9.6 241 22.4 134% Food shop 66 6.1 66 6.1 0% Take away shopping 6 0.6 10 0.9 67% Note: n=1076 Table 4. Online wine shopping during the lockdown. Online wine shopping in lockdown n. % Purchased wine online in lockdown 198 15.8 for the first time 86 43.4 as usual 66 33.3 more frequently 46 23.2 Notes: n=1076. 214 Bio-based and Applied Economics 10(3): 207-218, 2021 | e-ISSN 2280-6172 | DOI: 10.36253/bae-10044 Giulia Gastaldello, Daniele Mozzato, Luca Rossetto the closure of bars and ban on social gatherings, no sig- nificant effects emerge from drinking wine to socialize (R_SOC) as well as from drinking with friends and col- leagues before Covid (BC_FRICOL). Wine consumption frequency before the pandemic (WCONS_B) negatively impacts DV1 for regular consumers (i.e., drinking wine Table 5. LR on the decrease (DV1; Model A) and increase (DV2; Model B) of wine consumption frequency in lockdown.     A - Reduced wine consumption frequency in lockdown (DV1) B - Increased wine consumption frequency in lockdown (DV2) B S.E. Wald Exp(B) 95% C.I.for EXP(B) Sig. B S.E. Wald Exp(B)   95% C.I.for EXP(B) Sig. Lower Upper Lower Upper AGE 0.12 0.06 3.67 1.13 1.00 1.28 0.06  * -0.08 0.07 1.28 0.92 0.80 1.06 0.26   GENDER -0.13 0.17 0.57 0.88 0.63 1.23 0.45   -0.68 0.20 11.32 0.51 0.34 0.75 0.00 *** INCOME (good) 0.10   0.95   0.31   0.86   INCOME 1 (sufficient) 0.05 0.16 0.08 1.05 0.76 1.45 0.78   -0.01 0.19 0.00 0.99 0.68 1.44 0.96   INCOME (bad) 0.06 0.29 0.04 1.06 0.60 1.89 0.84   -0.19 0.34 0.31 0.83 0.43 1.61 0.58   CHILDY -0.44 0.17 6.71 0.65 0.46 0.90 0.01 ** 0.46 0.19 5.67 1.58 1.08 2.30 0.02 *** CRISFEAR -0.10 0.09 1.28 0.90 0.76 1.08 0.26   0.04 0.10 0.13 1.04 0.85 1.27 0.72   ISOFEEL 0.16 0.10 2.45 1.17 0.96 1.42 0.12   -0.14 0.12 1.52 0.87 0.69 1.09 0.22   VIRUSFEAR 0.08 0.09 0.92 1.08 0.92 1.28 0.34   -0.08 0.10 0.67 0.92 0.76 1.12 0.41   REFOFEEL -0.11 0.09 1.42 0.90 0.76 1.07 0.23   0.18 0.10 3.02 1.20 0.98 1.46 0.08 * VARBC_NO     61.74   0.00       46.82   0.00   VARBC_RED 1.28 0.19 44.28 3.59 2.47 5.24 0.00 *** 0.45 0.23 3.68 1.56 0.99 2.47 0.06 * VARBC_INCR -0.18 0.31 0.33 0.84 0.46 1.53 0.57   2.24 0.34 44.17 9.40 4.86 18.20 0.00 *** VARSP_NO     11.44   0.00       12.95   0.00   VARSPC_RED 0.52 0.19 7.91 1.68 1.17 2.41 0.01 *** 0.28 0.22 1.57 1.32 0.86 2.03 0.21   VARSPC_INCR -0.29 0.35 0.66 0.75 0.38 1.50 0.42   1.37 0.38 12.94 3.94 1.87 8.30 0.00 *** INCREXP_W -1.80 0.21 77.15 0.17 0.11 0.25 0.00 *** 2.77 0.21 169.74 15.92 10.50 24.13 0.00 *** INCREXP_B 0.36 0.27 1.72 1.43 0.84 2.44 0.19   -0.96 0.32 9.14 0.39 0.21 0.71 0.00 *** INCREXP_SP 0.33 0.38 0.76 1.40 0.66 2.96 0.39   -0.36 0.41 0.78 0.70 0.31 1.56 0.38   BL_ALONE 0.34 0.21 2.57 1.41 0.93 2.14 0.11   -0.19 0.25 0.62 0.83 0.51 1.33 0.43   BL_FAM 0.36 0.19 3.76 1.44 1.00 2.07 0.05 ** 0.13 0.21 0.38 1.14 0.75 1.72 0.54   BL_FRICOL -0.15 0.22 0.46 0.86 0.56 1.32 0.50   0.32 0.26 1.49 1.38 0.82 2.31 0.22   LC_ONLINE 0.39 0.24 2.59 1.47 0.92 2.36 0.11   -0.31 0.28 1.25 0.73 0.43 1.26 0.26   FREQ_DIGIMEET 0.02 0.08 0.04 1.02 0.86 1.20 0.84   -0.10 0.10 0.96 0.91 0.75 1.10 0.33   R_SOC -0.22 0.17 1.66 0.80 0.57 1.12 0.20   -0.09 0.20 0.20 0.92 0.62 1.36 0.66   R_TASTE -0.11 0.19 0.35 0.89 0.62 1.30 0.56   0.50 0.23 4.59 1.65 1.04 2.61 0.03 ** R_FOOD -0.11 0.18 0.41 0.89 0.64 1.26 0.52   0.18 0.21 0.73 1.19 0.80 1.79 0.39   R_HEALTH 0.81 0.26 10.22 2.26 1.37 3.72 0.00 *** -0.18 0.28 0.41 0.84 0.48 1.45 0.52   R_RELAX -0.43 0.19 4.84 0.65 0.45 0.96 0.03 ** 0.74 0.22 11.86 2.10 1.38 3.20 0.00 *** SCBL_SUPER 0.06 0.17 0.11 1.06 0.76 1.46 0.74   -0.13 0.20 0.44 0.88 0.60 1.29 0.51   SCBL_ONLINE 0.39 0.23 2.70 1.47 0.93 2.32 0.10 -0.11 0.27 0.17 0.90 0.53 1.51 0.68   SCBL_DSALE 0.15 0.17 0.79 1.16 0.84 1.61 0.37   0.01 0.20 0.00 1.01 0.69 1.48 0.98   SCBL_WSHOP -0.07 0.16 0.16 0.94 0.68 1.28 0.69   -0.18 0.19 0.93 0.84 0.58 1.21 0.34   WCONS_B (occasional)     30.26       0.00       24.37   0.00   WCONS_B (regular) -0.57 0.26 4.77 0.57 0.34 0.94 0.03 ** 0.85 0.32 7.20 2.33 1.26 4.34 0.01 ** WCONS_B (daily) -1.04 0.19 28.84 0.35 0.24 0.52 0.00 ** 1.15 0.23 24.21 3.16 2.00 5.01 0.00 *** LOC_SUPP -0.21 0.09 6.18 0.81 0.68 0.96 0.01 ** 0.22 0.10 4.65 1.24 1.02 1.51 0.03 ** Constant 0.30 0.75 0.15 1.34 0.69   -4.26 0.90 22.43 0.01 0.00   Notes: n=1018. * p < .09; ** p < .05; *** p < .001. A: Hosmer and Lemeshow Test: chi-square 10.24, sig. 0.249. Pseudo R-square: Cox & Snell=0.28, Nagelkerke=0.38. Accurancy= 75.6%. B: Hosmer and Lemeshow Test: chi-square 11.93, sig. 0.154. Pseudo R-square: Cox & Snell=0.36, Nagelkerke=0.50. Accurancy= 82.3%. 215Drinking Covid-19 away: wine consumption during the first lockdown in Italy Bio-based and Applied Economics 10(3): 207-218, 2021 | e-ISSN 2280-6172 | DOI: 10.36253/bae-10044 at least once a week, and daily wine drinkers). Moreo- ver, people who used to drink wine with family mem- bers before Covid (BC_FAM) show higher odds for a decreased consumption frequency. Finally, consumers willing to support Italian wine producers (LOC_SUPP) show significantly lower odds of drinking less frequently in lockdown. Factors driving an increased consumption frequen- cy in lockdown (model B; DV2) are having children (CHILDY), willingness to refocus on oneself (REFO- FEEL), drinking wine for relaxation (R_RELAX) and for its palatability (R_TASTE), and the willingness to support domestic wine producers. Indeed, all these pre- dictors are connected to greater odds of drinking wine more often in lockdown. Conversely, being female and spending more on beer decrease the odds of having con- sumed wine more often during the lockdown. Model B highlights a potential substitution effect in favor of wine. DV2 is significantly affected by increasing (VARBC_INCR) and decreasing (VARBC_RED) beer consumption frequency, although the former’s effect shows a greater magnitude. Crosstabulations (Table 6) reveal that 30.7% of interviewees who drink wine more often in lockdown have simultaneously reduced both beer and spirits consumption frequency, and the 11.2% has reduced beer drinking only. This substitution effect involves less than half of the sample, since 53.2% respondents have increased wine and at least one other alcoholic beverage. 5. DISCUSSION AND CONCLUSIONS Covid-19 has profoundly changed people’s lifestyle, disrupting everyday habits and exposing them to con- siderable psychological pressure. Our results highlight that such pressure arises primarily from concerns for the economic and financial uncertainty caused by the pan- demic, followed by the fear of the virus. Wine consumption is confirmed to be deeply rooted in the Italian population’s habits. Descriptive analysis reveals a major pre-existing preference for wine, which was the most assiduously consumed alcoholic bever- age before the lockdown. Despite the disruptive effect of the pandemic, our results highlight that the vast major- ity of the sample kept purchasing wine (75.7%) without lowering the average price per bottle (60.1%). Market data on agri-food products and supply during the pan- demic confirm a moderate but positive trend for wine sales during the lockdown (+9%), performing better than other beverages (+6%) (ISMEA, 2020a, 2020b, 2020c). Accordingly, most respondents kept consuming wine notwithstanding the substantial context changes. In this respect, both regular and daily wine drinkers, who are the most common consumers within the Italian popula- tion (ISTAT, 2020), are likely to have drunk wine more frequently during the lockdown rather than the reverse. Consistent with Wood et al. (2005), these findings rep- resent an indicator of strength of the wine consump- tion habit. Results also show that 22.5% of the sample stopped purchasing wine in lockdown, with a similar number of respondents (22.4%) who consumed wine from their wine stock. Although the current study did not investigate the size of this stock, this finding par- tially explains the non-significant association between a stop of wine shopping and reducing wine consumption in lockdown. It also reveals the presence of a wine stock in an interesting slice of Italian wine consumers, which calls for further investigations. Nevertheless, the impact of shock and habits disrup- tion emerges on sales channels, consumption occasions, and wine consumption frequency. Following the pan- demic, most consumers kept buying wine in supermar- kets, while mobility restrictions significantly penalized other important sales channels such as wine shops and direct sales. The online channel partly benefited from the lockdown, recording a 43% increase and managing to attract new buyers. Although the relevance of online wine sales is limited, these extraordinary circumstances may lead to long-run effects on this sale channel, accel- erating its growth in two ways: by pushing consumers to try online wine shopping, and by encouraging wine retailers to improve/create their online offer. This posi- tive trend is in line with market data, with the online demand for agri-food products recording a 141% growth during the two months of lockdown (IRI, 2020). Con- sumption occasions also suffered from the stringent lim- itations imposed, as the only options during a lockdown is consuming wine alone or with household members. Table 6. Substitution effect in favor of wine and beer during the lockdown.       Increased wine consumption frequency in lockdown No Yes Count % Count % Other 287 40.8% 175 53.2%