Bio-based and Applied Economics 9(2): 137-154, 2020 ISSN 2280-6180 (print) © Firenze University Press ISSN 2280-6172 (online) www.fupress.com/bae Full Research Article DOI: 10.13128/bae-8337 On the relevance of the Region-Of-Origin in consumers studies Fabio Gaetano Santeramo1, emilia lamonaca1,*, Domenico carlucci2, biaGia De DevitiiS1, antonio Seccia1, roSaria viScecchia1, Gianluca narDone1 1 University of Foggia (Italy) 2 University of Bari (Italy) Abstract. The existing literature on the consumers’ attitude toward region-of-origin (RoO) provides numerous and varying evidence on the relative importance of this extrinsic attribute as compared to other product characteristics. The article aims at characterising the heterogeneity in the relative importance of RoO. We systematical- ly review the literature on RoO and build an ad hoc indicator to measure the rela- tive importance of RoO as compared to other attributes of agri-food products under investigation. We then explain, through a meta-analytical approach, how the relative importance of RoO varies according to factors related to publication process, meth- odological issues, and characteristics of articles. Findings reveal the limited influ- ence of publication process and methodological issues on the relative importance of RoO. In contrast, we find a strong effect of characteristics of articles, with the relative importance of RoO being highly dependent on products and origins under investiga- tion. The results also highlight that RoO is an effective differentiation instrument in the agri-food markets only if supported by geographical indication labels. Managerial implications are also provided. Keywords. Agri-food, Consumer, Meta-analysis, Region-of-origin, Systematic review. JEL codes. Q13, P46, M31. 1. Introduction Regional imagery is increasingly being recognised as having a commercial value for agri-food products. It provides a subjective source of quality differentiation (Henchion and Mcintyre, 2000; Marcoz et al., 2016). In fact, even though countries operate within an increasingly globalised context, the indication of the region-of-origin (RoO) of agri-food products still appears to be a relevant cue for both consumers and producers or marketers (Pucci et al., 2017). The RoO of agri-food products still matters when examining consum- *Corresponding author. E-mail: emilia.lamonaca@unifg.it Editor: Davide Viaggi. 138 Fabio Gaetano Santeramo et al. ers’ product evaluations and buying behaviour (Chamorro et al., 2015). For producers and marketers, RoO allows them to charge prices above marginal cost, thus achieving market power. By using a regional indication, producers and marketers are able to exploit exist- ing associations consumers have with RoO and provide their product with an image (Bru- wer et al., 2012). Indeed, the strategic advantage of regional branding is that an agri-food product can be differentiated on the basis of geographic origin, an unique attribute dif- ficult to reproduce and presumed to be a quality cue for the product (van Ittersum et al., 2007; Chan and Marafa, 2013). The existing literature on the consumers’ attitude towards RoO provides numerous and varying evidences on the relative importance of this extrinsic attribute as compared to other product characteristics. The RoO effect has been analysed, among others, by Hen- chion and Mcintyre (2000), who concluded that RoO is an important consideration for two out of three Irish consumers when deciding to buy quality products and that prod- ucts from rural areas are generally perceived to be of high quality. In addition, Stefani et al. (2006) showed that the narrower and more precisely defined the RoO, the higher the quality expectation of consumers supporting the role of origin as a quality cue. Empiri- cal evidence shows that RoO effect on product evaluation is product-specific and varies depending on the characteristics of consumers. In particular, consumers from different countries tend to perceive RoO in a different manner and their knowledge influences the impact of RoO on their behaviour (Perrouty et al., 2006). In this regard, Engelbrecht et al. (2014) demonstrated that RoO of wine plays a secondary role in influencing consumers when faced with a purchasing decision on its own, while Dekhili and d’Hauteville (2009) found that the image of RoO has a specific influence on consumers’ selection behaviour for olive oil although with differences between consumers from different countries. Simi- larly, Dekhili et al. (2011) showed that French consumers tend to choose olive oil based on official signals, while Tunisian consumers mainly use RoO and sensory cues. Differ- ences across consumers emerges also at the regional level, as in Aranda et al. (2015) who found that Spanish consumers tend to value La Mancha region less than Rioja region in choosing wine. Their findings suggest different level of importance for RoO, relative to other products’ attribute under investigation. Overall, RoO has an effect if consumers per- ceive substantial differences between regions in terms of their product-origin associations (Marcoz et al., 2016). Literature on RoO is vast and fragmented, so we aim at characterising the heterogene- ity in the relative importance of RoO. On the basis of a systematic review of the literature on RoO, we have built an ad hoc indicator to measure the relative importance of RoO as compared to other attributes of agri-food products under investigation. We have then explained, through a meta-analytical approach, how the relative importance of RoO var- ies according to specific factors related to publication process, methodological issues, and characteristics of articles. The meta-analysis is based on a sample of 27 papers, which dif- fer by products and origins under investigation, and type of methodological framework. We have also expanded the study by Santeramo and Lamonaca (2020), who evaluated geographical label in consumers’ decision-making process, by proposing a quantile regres- sion analysis. The quantile regression allows us a better representation of the heterogeneity in the index measuring the relative importance of RoO. Indeed, the quantile regression estimator is robust, which means that the influence of outlying observations is bound. 139On the relevance of the Region-Of-Origin in consumers studies Our analysis would identify patterns in heterogeneous results in the vast body of research that examines the regional branding construct and the various effects of RoO on consumer buying behaviour (e.g. Atkin et al., 2017; Pucci et al., 2017). The success of regional branding strategies and of regulations protecting regional products largely depends on consumers’ evaluation of RoO that informs them on the authenticity of those products (van Ittersum et al., 2007). Our contribution is to provide a finer granular over- view of the RoO effects as relates to consumer product evaluations. A better understand- ing of consumers’ evaluation of RoO may benefit producers and marketers in designing differentiation strategies that support the competitiveness of regional products more effec- tively. Furthermore, it may facilitate policymakers in developing RoO-based communi- cation strategies and policies on the protection of regional products aimed at supporting rural economies, especially disadvantaged areas. The reminder of the article is as follows. The next section describes the protocol adopted for the systematic review of literature on consumers’ evaluation of RoO, as well as the quantitative methods used to examine determinants of heterogeneity in the rela- tive importance of RoO across studies. The results, presented in section 3, describe the contribution of publication process, methodological issues and characteristics of studies in explaining heterogeneity in the index measuring the relative importance of RoO as com- pared to other attributes of the product under investigation. The last section concludes with implications for the food industry and policymakers. 2. Materials and methods 2.1 Systematic review and sample description We systematically reviewed the literature based on RoO following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol (Moher et al., 2009; Shamseer et al., 2015). Our bibliographic research took place in July-September 2018 and focused on articles published in Scopus, including articles up to 20181. In order to be included in the quantitative synthesis (i.e. meta-analysis), papers had to meet two general criteria. First, they had to deal with consumers’ attitude, or preference, or intention to buy or willingness to pay for RoO of agri-food products. Second, papers had to provide a comparison between RoO and other products’ attributes. Further inclu- sion criteria allowed us to select only peer-reviewed published studies, supposed to be val- idated knowledge with a potentially greater impact in the field, and papers in English, the foremost language used to spread scientific knowledge. We identified an initial set of articles (n = 947) which contained all possible combina- tions of RoO-based, consumer-related, and sector-specific keywords in their title, abstract or keywords. In particular, we ran separate searches in Scopus using the following strings: [“place brand” OR “region-of-origin” OR “umbrella brand”] AND [“attitude” OR “attrib- ute” OR “behaviour” OR “choice” OR “consumer” OR “consumption” OR “preference” OR “segmentation” OR “willingness to pay”] AND [“agri food” OR “food”]. After removing duplicates (n = 680), two independent researchers screened 267 studies and selected eli- 1 We do not set time limits in the bibliographic search. 140 Fabio Gaetano Santeramo et al. gible articles on the basis of information contained in the full text (n = 60). We excluded 17 full text articles with reasons2. The steps of the systematic review are synthesised in a PRISMA flow diagram (figure 1). The final sample includes 43 articles, of which 27 (listed in table 1) included in the quantitative synthesis for a total of 194 observations (articles may include more than one observation). The vast majority of articles in the sample are published in peer-reviewed journals of high-medium prestige (48% in Q1, 37% in Q2); 37% of articles falls into the subject areas of Agricultural and Biological Sciences, 33% into Business, Management and Accounting, 26% into Economics, Econometrics and Finance. The first study was published in 2001 (van der Lans et al., 2001), however the interest for the topic has grown progressively overtime. Indeed, about two thirds of articles (74%) were published after 2010, demon- 2 A general review of place branding literature was excluded due to the lack of evidence on con- sumers’ attitude, or preference, or intention to buy or willingness to pay for RoO of agri-food prod- ucts. Other articles were excluded due to their focus, i.e. city brand not agri-food specific, private brands, physicochemical composition parameter and sensory attributes. Further articles not in Eng- lish (3 out of 17) and not peer-reviewed published (8 out of 17) were excluded. Figure 1. PRISMA flow diagram. Source: Elaboration on Moher et al. (2009). 141On the relevance of the Region-Of-Origin in consumers studies strating the dynamic character of the literature on RoO (figure 2). More than half of the papers share at least one of co-authors, and the author with the most articles on the topic is S. Mueller Loose (author of 3 articles). 2.2 A measure of the relative importance of RoO We identified 47 types of attributes within our sample, other than RoO. We ideally classified them in three categories (table 2): i) intrinsic attributes (13 types) which directly describe products; ii) extrinsic attributes (14 types) which indirectly characterise products (Dekhili and d’Hauteville, 2009); iii) additional attributes (20 types) which refer to the level of product knowledge and involvement (Arancibia et al., 2015). The most frequent attributes in our sample are two extrinsic attributes, namely price and packaging, followed by two intrinsic attributes, namely type and variety. In order to measure the relative importance of RoO as compared to other attributes of a product under investigation, we built an index specific for each i-th observation (i.e. estimate) within the j-th reviewed study. The index ( ) is equal to the ratio between the sum of the relative importance of each k-th attribute ( ) and the number of attrib- utes other than RoO (Ki,j): (1) where , the relative importance of the k-th attribute compared to RoO, may assume the following values: Figure 2. Evolution of the literature on Region-of-Origin. 0 5 10 15 20 25 30 0 1 2 3 4 5 cu m ul at iv e nu m be r ye ar -b y- ye ar n um be r Year-by-year Cumulative 142 Fabio Gaetano Santeramo et al. (2) The index, , measures the relative importance of RoO with respect to other generic attributes of a product, and ranges between -1 and 1: the higher the index, the Table 1. List of articles included in the meta-analysis. Reference Journal Journal rank Journal subject area Arancibia et al. (2015) Agricultural Economics Review Q2 ABS Atkin et al. (2017) Wine Economics and Policy Q1 EEF Bernabéu et al. (2012) British Food Journal Q1 BMA Bruwer et al. (2012) Journal of Foodservice Business Research Q3 ABS Bryła (2015) Appetite Q1 Nursing Dekhili and d’Hauteville (2009) Food Quality and Preferences Q1 ABS Dekhili et al. (2011) Food Quality and Preferences Q1 ABS Fernandes-Ferreira-Madureira et al. (2013) International Journal of Wine Business Research Q2 BMA Grebitus et al. (2018) Agribusiness Q2 EEF Hollebeek et al. (2007) Food Quality and Preferences Q1 ABS Johnson and Bruwer (2007) International Journal of Wine Business Research Q3 BMA Marcoz et al. (2016) International Journal of Tourism Research Q1 BMA McCutcheon et al. (2009) International Journal of Wine Business Research Q2 BMA Mtimet et al. (2013) Journal of International Food and Agribusiness Marketing Q2 BMA Mueller Loose and Szolnoki (2012) Food Quality and Preferences Q1 ABS Mueller Loose et al. (2013) Food Quality and Preferences Q1 ABS Mueller Loose and Szolnoki (2010) Food Quality and Preferences Q1 ABS Nunes et al. (2016) Wine Economics and Policy Q1 EEF Perrouty et al. (2006) Agribusiness Q2 EEF Rahnama and Fadei (2017) Journal of Food Products Marketing Q2 BMA Resano-Ezcaray et al. (2012) Food Policy Q1 EEF Robertson et al. (2018) Journal of Wine Research Q3 ABS Sanjuán-López et al. (2009) Spanish Journal of Agricultural Research Q3 ABS Scarpa et al. (2015) Agribusiness Q2 EEF Schnettler et al. (2018) British Food Journal Q1 BMA Sutanonpaiboon and Atkin (2012) Journal of Food Products Marketing Q2 BMA van der Lans et al. (2001) European Review of Agricultural Economics Q2 EEF Notes: The rank, provided by the Scimago Journal & Country Rank (SJR), refers to the date of publi- cation for the corresponding SJR subject area. Abbreviations are Agricultural and Biological Sciences (ABS), Business, Management and Accounting (BMA), Economics, Econometrics and Finance (EEF). 143On the relevance of the Region-Of-Origin in consumers studies greater the relative importance of RoO as compared to other attributes. The index is dis- tributed with mean 0.34 and standard deviation 0.64, however the relative importance of RoO tends to vary according to structural and methodological differences across studies. 2.3 Meta-analytical approach and data description We adopt a meta-regression approach to investigate the determinants of heterogene- ity in the relative importance of RoO as compared to other attributes of products under investigation. We regress the index measuring the relative importance of RoO ( ) on its accuracy (i.e. sample size3) and on a set of - and -type moderator variables: (3) 3 Stanley et al. (2008) suggest to use degrees of freedom (or sample size) as a measure of the accu- racy of the variable under investigation. Other meta-analyses on the issue follow the same approach (e.g. Deselnicup et al., 2013). Table 2. Relative frequencies (RF) of intrinsic, extrinsic, and additional attributes in the sample. Intrinsic RF Extrinsic RF Additional RF Type 0.53 Price 1.00 Distribution channel 0.39 Variety 0.45 Packaging 0.62 Frequency of choice 0.28 Appearance 0.26 Brand name 0.39 Accompaniment 0.23 Alcohol content 0.25 Appellation 0.33 Concerns for environment 0.21 Colour 0.23 Medal 0.32 Recommendation by others 0.16 Vintage 0.21 Label information 0.21 Concerns for health 0.07 Taste 0.20 Producer 0.20 Preparation format 0.07 Sensorial characteristics 0.15 Production process 0.14 Availability 0.06 Serving temperature 0.13 Organic label 0.13 Advertising 0.05 Hedonic liking 0.10 Country-of-origin 0.10 Consumption for specials 0.04 Product quality 0.09 Retailer 0.09 Nostalgia 0.04 Expiry date 0.03 Informed liking 0.06 Concerns for animal welfare 0.03 Smell 0.03 Concerns for safety 0.03 Curiosity 0.03 State 0.02 Fashion of consumption 0.03 Knowledge level 0.03 Loyalty 0.03 Pleasure of consumption 0.03 Touristic issues 0.03 Traceability 0.03 Uniqueness 0.03 Notes: The relative frequencies are computed on a total of 194 observations. 144 Fabio Gaetano Santeramo et al. where the accuracy of the index (Ni,j) models and corrects publication selection bias4; Χr is a vector of R regressors thought to affect the magnitude of the publication selection bias5; Ψs is a vector of S regressors, related to relevant characteristics of a study, that influence the magnitude of the index and explain its systematic variation across the observations (i) of the reviewed studies (j); γr and δs are coefficients which reflect the biasing effect of publication selection and of study’s characteristics; εi,j is an independently and identically distributed error term. The vector of X-type moderator variables includes information related to the publica- tion process and methodological issues (table 3). It controls, through a dummy variable, for the presence of more than one article published by the same author (42% of obser- vations). In order to account for the prestige of the journal, specific dummies control for articles published in Q1 (51% of observations), Q2 (baseline), and Q3 (6% of obser- vations) journals, according to the rank provided by the Scimago Journal & Country Rank at the date of publication. The dynamic character of the literature on the issue is accounted for using a dummy that discriminates between articles published before and after 20106, whereas a numerical variable controls for the cumulative number of articles published overtime. As for methodological issues, dummies control for methods and reference variables adopted to assess the relative importance of RoO. In our sample, we observe articles based on best-worst scaling analyses (6% of observations), choice mod- els (8% of observations), conjoint analyses (19% of observations), focus groups (1% of observations), hedonic price models (9% of observations), latent classes analyses (12% of observations), descriptive statistics (baseline). The reference variables mostly used are per- centages in terms of importance of attributes (41% of observations), average importance of attributes (40% of observations), estimated willingness to pay (3% of observations), beta (baseline). The set of Ψ-type moderator variables includes dummies related to specific characteristics of studies, to account for heterogeneity in the relative importance of RoO (table 3). It controls for specific product category such as olive oil (13% of observations), wine (58% of observations), other products (baseline), and origin such as Argentina (13% of observations), Australia (14% of observations), Chile (2% of observations), New Zea- land (5% of observations), Tunisia (4% of observations), United States (9% of observa- tions), other countries (baseline). Lastly, a dummy identifies paper that associate a certi- fied label, such as PDO or PGI, to RoO (25% of observations). In order to correct for heteroskedasticity and to obtain efficient estimates, we normal- ised all but one elements (i.e. Xr7) of the equation (3) by the accuracy of the index, Ni,j. 4 Publication selection may distort evidence from literature, undermining the external validity of inferences and implications (Santeramo and Lamonaca, 2021). Biases from publication selection may occur if certain results are more likely to be published (e.g. statistical significant results, esti- mated coefficients of certain sign or magnitude) (Stanley, 2005). 5 The X-type moderator variables allow us to capture the wide dimension of selection bias, which is a complex socio-economic phenomenon that goes beyond the mere publication selection (Stanley et al., 2008). 6 The year 2010 is the median year of the articles in the sample. In addition, 72% of observations are included in articles post 2010. 7 The rationale of the exclusion is that Χ-type moderator variables may influence the likelihood of acceptance for publication, but should not be informative on the index. 145On the relevance of the Region-Of-Origin in consumers studies After the normalisation, the intercept and slope coefficients are reversed from the equa- tion (3). The new intercept (α0) is a test for publication selection bias and the new slope (α) is a test for the average value beyond the publication selection bias (Stanley et al., 2008). If statistically significant, α0 suggests the existence of publication selection bias, and α allows to conclude on the accuracy of the index (Santeramo and Shabnam, 2015). We used Probit specifications to assess how determinants of heterogeneity in the relative importance of RoO influence the likelihoods of observing lower or higher values of the index . These likelihoods are captured by two dependent variables defined as dummies, that distinguish between cases in which RoO tend to be less or more important for consumers as compared to other attributes of the product under investigation. The likelihood of observing lower values of the index equals to 1 for negative observations of , and zero otherwise. The likelihood of observing higher values of the index equals to 1 for positive observations of , and zero otherwise. Observations of equal to zero, indicating that RoO is impor- tant as much as a generic attribute of the product under investigation, serve as baseline. Table 3. List and description of moderator variables. Moderator variable Type of variable Mean Std. dev. Obs. X-type moderator variables Sample size Numerical 372.89 373.47 159 Authorship Dummy 0.42 0.50 194 Q1 (journal prestige) Dummy 0.51 0.50 194 Q3 (journal prestige) Dummy 0.06 0.24 194 Post-2010 Dummy 0.72 0.45 194 Number of paper (cumulative) Numerical 15.41 7.07 194 Best-worst scaling analysis (method) Dummy 0.06 0.24 194 Choice model (method) Dummy 0.08 0.27 194 Conjoint analysis (method) Dummy 0.19 0.39 194 Focus group (method) Dummy 0.01 0.07 194 Hedonic price model (method) Dummy 0.09 0.29 194 Latent classes analysis (method) Dummy 0.12 0.33 194 % (reference variable) Dummy 0.41 0.49 194 Avg. (reference variable) Dummy 0.40 0.49 194 WTP (reference variable) Dummy 0.03 0.16 194 Ψ-type moderator variables Certified origin Dummy 0.25 0.43 194 Argentina Dummy 0.13 0.34 194 Australia Dummy 0.14 0.35 194 Chile Dummy 0.02 0.14 194 New Zealand Dummy 0.05 0.22 194 Tunisia Dummy 0.04 0.20 194 United States Dummy 0.09 0.28 194 Olive oil Dummy 0.13 0.34 194 Wine Dummy 0.58 0.50 194 146 Fabio Gaetano Santeramo et al. The model in equation (3) is also estimated in a quantile regression fashion. These models allow us to identify factors determining more or less importance of RoO (observa- tions of within 50th percentile), less importance of RoO (observations of within 25th percentile), more importance of RoO (observations of within 75th percentile). The quantile regression allows us to particularise the dependency of the index on deter- minant of heterogeneity in the relative importance of RoO for every quantile and, thus, it can be tailored to the extremes by conditioning on lower quantiles. In addition, the quan- tile regression estimator is robust, which means that the influence of outlying observations is bound. These properties lead to a better representation of the heterogeneity in the index . 3. Results and discussion The results of the Probit and quantile regression models are presented in table 4. The analysis of the constant term (α0) and slope coefficient (α) allows us to detect potential publication selection bias. Looking at the results from Probit models, we find that publica- tion selection bias is more likely to occur for higher values of the index, measuring the relative importance of RoO as compared to other product’s attributes under investigation; in contrast, lower values of the index are less likely to be affected by publication selection bias. However, the results of the quantile regression reveal that the constant term is not significantly different from zero at any conventional level, suggesting that the publication selection bias is not a major issue in our sample. We conclude that publication selection does not distort evidence from the literature on RoO, or undermine the external valid- ity of inferences and implications on the relative importance of RoO (Stanley, 2005). This implies that all the heterogeneity we observe in the index depends on publication pro- cess, methodological issues, and characteristics of studies. Given the absence of publica- tion selection bias, values of the index within the 50th and 75th percentiles tend to be more accurate with the estimated coefficients being positive and significant at 1% level. Reflecting on the publication process, the likelihood of having RoO more impor- tant increases for studies co-authored by experienced scholars (column A); in addition, the higher the values of the index (75th percentile), the greater the importance of RoO (column E). As for journal prestige, the quantile regression results show that the relative importance of RoO decreases in articles published in Q3 journals. This is true in particu- lar for higher values of the index (within 50th and 75th percentile). A comparable result is found in Probit models, where the coefficients estimated for articles published in Q1 and Q3 journals are negative and significant in the specifications in column (B). The probabil- ity of having RoO more important decreases for studies published in medium-high pres- tigious journals. However, the probability of having RoO less important increases for stud- ies published in journals with lower prestige (Q3, for which the estimated coefficient in the specification in column (A) is positive and significant). It is worth noting that, in our sample, the vast majority of articles published in Q3 journals are wine-based studies (e.g. Johnson and Bruwer, 2007; Bruwer et al., 2012; Robertson et al., 2018) that tends to be negatively correlated with the relative importance of RoO (negative and significant coef- ficient reported in column E). Overall, the relative importance of RoO tends to decrease with the prestige of the journal in which articles are published. Similarly, Santeramo and 147On the relevance of the Region-Of-Origin in consumers studies Lamonaca (2019) argue that the authorship and the prestige of the publication outlet help in explaining the variability in studies’ outcome. If articles are published after 2010, RoO less important is less likely to be observed; the opposite is true for RoO more important. This evidence support the idea that the importance of RoO is likely to increase when the background on the issue is based on a wider set of empirical evidence; indeed, evidences available on the issue show an increasing trend overtime (cfr. figure 2) and the vast major- ity of them, in our sample, are observed in articles published after 2010. As for methodological issues, both methods or reference variables used in our sam- ple tend to have a limited influence in determining the relative importance of RoO. The few exception are best-worst scaling and latent classes analyses. The former tends to be associated with higher relative importance of RoO (positive and significant coefficient reported in column E); the latter tends to provide evidence of lower likelihood of RoO more important (negative and significant coefficient reported in column B). Our results differ from Deselnicup et al. (2013), who find a positive influence of conjoint analysis and hedonic price model on the price premium for origin-based labels (not statistically sig- nificant in our specifications). The result is not surprising. In fact, estimation techniques may influence the estimated willingness to pay (WTP), whereas latent characteristics of specific sub-sample within the population under investigation are likely to affect the rela- tive importance of RoO. The characteristics of studies have a varying contribution on the relative importance of RoO. The relative importance of RoO decreases in country-specific studies (columns C-E), in particular for Australia and New Zealand for which the probability of having RoO less important increases (column A) and the probability of having RoO more impor- tant decreases (column B). As suggested in Dekhili et al. (2011), nationality or culture appears to influence consumers’ perceptions of the importance of RoO as compared to other product’s attributes. Similarly, Perrouty et al. (2006) show that consumers from dif- ferent countries tend to perceive RoO in a different manner and their knowledge influ- ences the impact of RoO on their behaviour. Verbeke et al. (2012) suggest the existence of substantial differences in consumers’ awareness of geographical origin between coun- tries with versus countries without a tradition of geographical indications in their agri- food quality policies. For instance, such awareness tends to be higher in countries with a strong tradition of using these quality schemes, e.g. Southern (Italy and Spain) and West- ern (France) Europe. We also find product-specific differences; the relative importance of RoO decreases in the analysis of olive oil (negative and significant coefficients in col- umns C-E), for which the probability of having RoO more important is decreasing (nega- tive and significant coefficient in column B). In fact, consumers tend to judge olive oil more for intrinsic than for extrinsic attributes (Dekhili and d’Hauteville, 2009; Dekhili et al., 2011). Lastly, the relative importance of RoO increases in studies where RoO is certi- fied by an origin-based label. Our result is in line with findings from Deselnicup et al. (2013), who provide evidence of greater price premiums for product with certified ori- gin than one using a non-regulated regional name. The interest in the origin of foods is a strong direct and indirect driver of consumers’ use of labels (Verbeke et al., 2012). Add- ing regional certification labels (e.g. Protected Designation of Origin –DOP–, Protected Geographical Indication –PGI–, American Viticultural Area–AVA–) allows to strengthen regional branding, in particular in the case of lesser known regions (Bruwer and Johnson, 148 Fabio Gaetano Santeramo et al. Table 4. Probit model estimation: analysis of publication selection bias. Variables Probit estimates Quantile regression estimates Negative index (A) Positive index (B) 25th percentile (C) 50th percentile (D) 75th percentile (E) Constant (α0) -1.291* 2.215*** 0.001 -0.001 -0.001 (0.729) (0.841) (0.004) (0.00325) (0.002) Bias (α0) 6.324 -21.900 -0.106 0.573*** 0.591*** (50.360) (50.200) (0.207) (0.173) (0.087) Authorship -0.845 1.892** 0.001 0.002 0.002** (0.597) (0.762) (0.003) (0.002) (0.001) Q1 (journal prestige) 0.637 -1.061** -0.001 -0.001 -0.001 (0.577) (0.511) (0.003) (0.002) (0.001) Q3 (journal prestige) 1.268* -1.451* -0.001 -0.012*** -0.004*** (0.705) (0.873) (0.003) (0.003) (0.001) Post-2010 -2.627*** 4.590*** 0.004 0.004 0.002 (0.836) (1.094) (0.004) (0.003) (0.002) Cumulative 0.146** -0.235*** -0.0003 -0.0001 -0.00004 (0.058) (0.073) (0.0003) (0.0002) (0.0001) Best-worst (method) Omitted Omitted 0.003 0.001 0.004** (0.004) (0.003) (0.002) Choice (method) -0.798 -0.439 -0.0001 0.001 0.001 (0.652) (0.660) (0.003) (0.003) (0.001) Conjoint (method) 0.348 -0.358 -0.001 -0.001 0.001 (0.501) (0.503) (0.003) (0.002) (0.001) Focus group (method) Omitted Omitted 0.003 0.003 0.003 (0.008) (0.007) (0.004) Hedonic price (method) 0.192 -1.163 -0.001 -0.0003 -0.001 (0.722) (0.765) (0.003) (0.003) (0.001) Latent class (method) 0.816 -2.021*** -0.001 -0.002 -0.0002 (0.689) (0.655) (0.003) (0.003) (0.001) % (reference variable) -0.479 0.254 0.002 0.001 0.001 (0.361) (0.332) (0.002) (0.002) (0.001) Avg. (reference variable) -0.662 -0.598 0.002 0.001 0.001 (0.567) (0.650) (0.003) (0.003) (0.001) WTP (reference variable) 0.748 Omitted -0.001 -0.001 -0.003* (0.498) (0.004) (0.004) (0.002) Certified origin -0.327 0.330 0.440** 0.431*** 0.409*** (0.289) (0.376) (0.187) (0.156) (0.079) Argentina 4.470 -4.236 -0.447 -0.106 0.250 (3.187) (3.338) (1.469) (1.228) (0.618) Australia 2.616*** -3.562*** -0.749*** -0.689*** -0.609*** (0.730) (1.001) (0.253) (0.212) (0.107) Chile Omitted 1.642* 0.702* -0.004 -0.123 (0.929) (0.396) (0.331) (0.167) 149On the relevance of the Region-Of-Origin in consumers studies 2010). Similarly, van der Lans et al. (2001) find that the RoO cue and PDO label influence regional product preferences through perceived quality. In a nutshell, the relative importance of RoO is highly dependent on structural char- acteristics of studies and, to a lower extent, on issues related to the publication process and methodological issues. The paper however is not exempt from limitations. The evaluation of the relative importance of RoO for consumers, through a meta-analytical approach, is based on infor- mation retrieved from literature, thus, it is highly dependent on the quality of each article. Although the comprehensive analysis of heterogeneity in the relative importance of RoO should minimise the biasing effect due to the quality of each article, further studies on the issue should consider to applying a quality assessment tool of articles included in the quantitative synthesis (e.g. Cox et al., 2016). 4. Conclusions and implications The existing literature on the consumers’ attitude toward region-of-origin (RoO) pro- vides numerous and varying evidences on the relative importance of this extrinsic attrib- ute as compared to other product characteristics. In order to characterise the heteroge- neity in the relative importance of RoO, we systematically reviewed a large number of Variables Probit estimates Quantile regression estimates Negative index (A) Positive index (B) 25th percentile (C) 50th percentile (D) 75th percentile (E) New Zealand 3.037** -4.326*** -0.762** -1.310*** -1.331*** (1.253) (1.291) (0.342) (0.286) (0.144) Tunisia 2.164 Omitted 0.960 0.154 -0.093 (2.074) (0.695) (0.581) (0.292) United States 5.237 -11.470** -0.915 -0.914 -0.043 (5.005) (4.921) (1.912) (1.598) (0.805) Olive oil 0.887 -1.733** -0.950*** -0.825*** -0.379** (0.612) (0.772) (0.359) (0.300) (0.151) Wine -0.323 0.823 -0.132 -0.066 -0.216** (0.524) (0.577) (0.242) (0.202) (0.102) Observations 142 137 159 159 159 Notes: Probit and quantile regression estimates of model in equation (3). The dependent variable is a dummy equal to 1 for negative observations of the index in specification in column (A), a dummy equal to 1 for positive observations of the index in specification in column (B), the index in specifica- tions in columns (C), (D) and (E). Coefficients estimated for -type moderator variables, related to study characteristics, have been scaled by a factor of 102 in specifications in columns (A) and (B). The index is -0.001 in 25th percentile, 0.001 in 50th percentile, 0.003 in 75th percentile. Omitted variables in Probit models due to a perfect prediction of failure for observations different from zero. *** Significant at the 1 percent level. ** Significant at the 5 percent level. * Significant at the 10 percent level. 150 Fabio Gaetano Santeramo et al. studies on the issue and provided a quantitative synthesis of empirical evidences on the consumers’ perception of RoO. We explained the differences in the relative importance of RoO with several control factors related to publication process, methodological issues, and characteristics of articles. The meta-regression results allowed us to conclude on the limited influence of publi- cation process and methodological issues on the relative importance of RoO. In contrast, we found a strong effect of characteristics of articles, with the relative importance of RoO being highly dependent on products and origins under investigation. We can also con- clude that RoO is an effective differentiation tool in the agri-food markets only if support- ed by geographical indication (GI) labels, such as Protected Designation of Origin (PDO), Protected Geographical Indication (PGI), American Viticultural Area (AVA). For instance, it is well-known the higher propensity of consumers in attributing a great importance to GI labels for agri-food products; consumers benefit from GI schemes that certify qual- ity at different geographical levels (van Ittersum et al., 2007; Verbeke et al., 2012). In this regard, it is worth of mention the positive relation between RoO and GI; the addition of regional information on a product label increases consumer confidence in the quality of that product (Bruwer and Jhonson, 2010). Overall, our study suggests that protecting and marketing agri-food products with regional certification labels, such as PDO or PGI, may be beneficial for producers and mar- keters. They should fine-tune the differentiation of agri-food products through RoO, par- ticularly when RoO have a positive reputation (Santeramo et al., 2020a, b). Consider as a representative example the Tuscan Experience, characterised by a strong regional image (Stefani et al., 2006; Bryła, 2015). It is therefore critical for policymakers to develop focused communication strategies towards consumers in order to convey attractive information about RoO that, as suggested in Verbeke et al. (2012), stimulates their interest in the origin of foods and builds favourable perceptions about quality and distinctiveness of products with RoO labels. For instance, Italian consumers stated that the label of origin “produced in Puglia” is considered the preferred attribute for mozzarella cheese due to the high repu- tation of this region for mozzarella production (Viscecchia et al., 2019). The effectiveness of communication strategies should be enhanced by targeting different messages to differ- ent target markets (van Ittersum et al., 2007; Marcoz et al., 2016) and by a new concept of label in terms of contents and communication channels (Corallo et al., 2019). Indeed, our analysis revealed that the importance of RoO for consumers tends to vary according to products and countries involved. Hence, communication efforts should stimulate con- sumers’ interest in RoO, especially for wine and in countries without a strong tradition of geographical indications in their agri-food quality policies. Furthermore, policymakers should consider the benefits of a collaborative marketing program for regional products. Indeed, while many regional products are already under regional certification labels, many more remain out of the protection of an incisive regional logo. In this regard, examples of best practices come from Rural Development Programmes 2014-2020 implemented in the EU Member States, where the Measure 03 “Quality schemes for agricultural products and foodstuffs” allows local policymakers to support regional agri-food products in order to improve competitiveness of producers, create value added for agri-food products of high quality, promote regional products at the local, national and international level. The Meas- ure 03 also compensate producers for costs arising from specific management activities 151On the relevance of the Region-Of-Origin in consumers studies required to adhere to quality schemes. Similar policy approaches would benefit consumers, who obtain information on the authenticity of regional products, producers, who enhance competitiveness in marketing regional products, and overall rural economies, in particu- lar disadvantaged areas. As suggested in Atkin et al. (2017), a cohesive effort in promoting strong regional labels may result in growth and success. References Arancibia, R.G., Rossini, G.and Guiguet, E.D. (2015). Wine label descriptors and shelf price paid by argentine consumers. Agricultural Economics Review 16(2): 56-72. Aranda, E., Gómez, M. and Molina, A. (2015). Consumers’ brand images of wines: Differ- ences between two leading spanish denominations of origin. British Food Journal 117(8): 2057-2077. Atkin, T., Wilson, D., Thach, L. and Olsen, J. (2017). Analyzing the impact of conjunctive labeling as part of a regional wine branding strategy. Wine Economics and Policy 6(2): 155-164. Belsley, D. A., Kuh, E., Welsch, R.E. Regression Diagnostics. New York: Wiley, 1980. Bernabéu, R., Díaz, M., Olivas, R., Olmeda, M. (2012). Consumer preferences for wine apply- ing best-worst scaling: A spanish case study. British Food Journal, 114(9), 1228-1250. Bruwer, J., Johnson, R. (2010). Place-based marketing and regional branding strategy per- spectives in the California wine industry. Journal of Consumer Marketing, 27(1), 5-16. Bruwer, J., Li, E., Lang, S., Chaumont, S. (2012). Perspectives on consumers’ attitudes to wine’s region of origin in a restaurant setting. Journal of Foodservice Business Research, 15(3), 300-318. Bryła, P. (2015). The role of appeals to tradition in origin food marketing. A survey among polish consumers. Appetite, 91, 302-310. Chamorro, A., Rubio, S. Miranda, F.J. (2015). The region-of-origin (ROO) effect on pur- chasing preferences: the case of a multiregional designation of origin. British Food Journal, 117(2), 820–839. Chan, C.S., Marafa, L.M. (2013). A review of place branding methodologies in the new millennium. Place Branding and Public Diplomacy, 9(4), 236-253. Corallo, A., Latino, M.E., Menegoli, M., De Devitiis, B. and Viscecchia, R. (2019). Human Factor in Food Label Design to Support Consumer Healthcare and Safety: A Sys- tematic Literature Review. Sustainability 11(15): 4019. Cox, D.N., Hendrie, G.A., Carty, D. (2016). Sensitivity, hedonics and preferences for basic tastes and fat amongst adults and children of differing weight status: A comprehen- sive review. Food Quality and Preference, 48, 359-367. Dekhili, S., d’Hauteville, F. (2009). Effect of the region of origin on the perceived quality of olive oil: An experimental approach using a control group. Food Quality and Pref- erence, 20(7), 525-532. Dekhili, S., Sirieix, L., Cohen, E. (2011). How consumers choose olive oil: The importance of origin cues. Food Quality and Preference, 22(8), 757-762. Deselnicu, O.C., Costanigro, M., Souza-Monteiro, D.M., McFadden, D.T. (2013). A meta- analysis of geographical indication food valuation studies: what drives the premium for origin-based labels?. Journal of Agricultural and Resource Economics, 204-219. 152 Fabio Gaetano Santeramo et al. Dias, C., Mendes, L. (2018). Protected Designation of Origin (PDO), Protected Geograph- ical Indication (PGI) and Traditional Speciality Guaranteed (TSG): A bibiliometric analysis. Food Research International, 103, 492-508. Engelbrecht, J.A., Herbst, F. and Bruwer, J. (2014). Region-of-origin (ROO) certification as marketing strategy in the South African wine market. International Journal of Wine Business Research 26(2): 139-162. Fernandes Ferreira Madureira, T.C. and Simões de Sousa Nunes, F.J. (2013). Relevant attributes of portuguese wines: Matching regions and consumer’s involvement level. International Journal of Wine Business Research 25(1): 75-86. Grebitus, C., Peschel, A.O. and Hughner, R.S. (2018). Voluntary food labeling: The addi- tive effect of “free from” labels and region of origin. Agribusiness 1-14. Henchion, M. and Mcintyre, B. (2000). Regional imagery and quality products: The Irish experience. British Food Journal 102(8): 630-644. Hollebeek, L.D., Jaeger, S.R., Brodie, R.J. and Balemi, A. (2007). The influence of involve- ment on purchase intention for new world wine. Food Quality and Preference 18(8): 1033-1049. Johnson, R. and Bruwer, J. (2007). Regional brand image and perceived wine quality: The consumer perspective. International Journal of Wine Business Research 19(4): 276- 297. Marcoz, E.M., Melewar, T.C. and Dennis, C. (2016). The value of region of origin, pro- ducer and protected designation of origin label for visitors and locals: The case of Fontina cheese in Italy. International Journal of Tourism Research 18(3): 236-250. McCutcheon, E., Bruwer, J. and Li, E. (2009). Region of origin and its importance among choice factors in the wine-buying decision making of consumers. International Jour- nal of Wine Business Research 21(3): 212-234. Moher, D., Liberati, A., Tetzlaff, J. and Altman, D.G. (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): e1000097. Mtimet, N., Zaibet, L., Zairi, C. and Hzami, H. (2013). Marketing olive oil products in the Tunisian local market: The importance of quality attributes and consumers’ behav- ior. Journal of International Food and Agribusiness Marketing 25(2): 134-145. Mueller Loose, S., Peschel, A. and Grebitus, C. (2013). Quantifying effects of convenience and product packaging on consumer preferences and market share of seafood prod- ucts: The case of oysters. Food Quality and Preference 28(2): 492-504. Mueller Loose, S. and Szolnoki, G. (2010). The relative influence of packaging, labelling, branding and sensory attributes on liking and purchase intent: Consumers differ in their responsiveness. Food Quality and Preference 21(7): 774-783. Mueller Loose, S. and Szolnoki, G. (2012). Market price differentials for food packaging characteristics. Food Quality and Preference 25(2): 171-182. Nunes, F., Madureira, T., Oliveira, J.V. and Madureira, H. (2016). The consumer trail: Applying best-worst scaling to classical wine attributes. Wine Economics and Policy 5(2): 78-86. Perrouty, J.P., D’Hauteville, F. and Lockshin, L. (2006). The influence of wine attributes on region of origin equity: An analysis of the moderating effect of consumer’s perceived expertise. Agribusiness 22(3): 323-341. 153On the relevance of the Region-Of-Origin in consumers studies Pucci, T., Casprini, E., Rabino, S. and Zanni, L. (2017). Place branding-exploring knowl- edge and positioning choices across national boundaries: The case of an Italian superbrand wine. British Food Journal 119(8): 1915-1932. Rahnama, H. and Fadaei, M. (2017). Modeling the factors affecting organic or local cucumber purchasing motivations in Northern Iran. Journal of Food Products Mar- keting 23(6): 705-716. Resano, H., Sanjuán, A.I. and Albisu, L.M. (2012). Consumers’ response to the EU quality policy allowing for heterogeneous preferences. Food Policy 37(4): 355-365. Robertson, J., Ferreira, C. and Botha, E. (2018). The influence of product knowledge on the relative importance of extrinsic product attributes of wine. Journal of Wine Research 29(3): 159-176. Sanjuán-López, A.I., Resano-Ezcaray, H. and Camarena-Gómez, D.M. (2009). Developing marketing strategies for jiloca saffron: A price hedonic model. Spanish Journal of Agricultural Research 7(2): 305-314. Santeramo, F.G. and Lamonaca, E. (2019). The Effects of Non‐tariff Measures on Agri‐ food Trade: A Review and Meta‐analysis of Empirical Evidence. Journal of Agricul- tural Economics 71(3): 595–617. Santeramo, F.G. and Lamonaca, E. (2020). Evaluation of geographical label in consum- ers’ decision-making process: a systematic review and meta-analysis. Food Research International 108995. Santeramo, F.G. and Lamonaca, E. (2021). Objective risk and subjective risk: The role of information in food supply chains. Food Research International 139: 109962. Santeramo F.G., Lamonaca, E., De Devitiis, B., Viscecchia, R., Nardone, G. (2020a). Assessing the role of regional labels for consumers. Acta Horticolturae [In press] Santeramo F.G., Lamonaca, E., Carlucci, D., De Devitiis, B., Seccia, A., Viscecchia, R., Nar- done, G. (2020b). Innovative quality-based strategies for the agri-food sector. Acta Horticolturae [In press] Santeramo, F.G. and Shabnam, N. (2015). The income-elasticity of calories, macro-and micro- nutrients: What is the literature telling us?. Food Research International 76: 932-937. Scarpa, R., Philippidis, G. and Spalatro, F. (2005). Product-country images and preference heterogeneity for mediterranean food products: A discrete choice framework. Agri- business 21(3): 329-349. Schnettler, B., Sepúlveda, N., Bravo, S., Grunert, K.G. and Hueche, C. (2018). Consumer acceptance of a functional processed meat product made with different meat sourc- es. British Food Journal 120(2): 424-440. Shamseer, L., Moher, D., Clarke, M., Ghersi, D., Liberati, A., Petticrew, M., et al. (2015). Preferred reporting items for systematic review and meta-analysis protocols (pris- map) 2015: Elaboration and explanation. BMJ 349: g7647. Stanley, T.D. (2005). Beyond publication bias. Journal of Economic Surveys 19(3): 309-345. Stanley, T.D., Doucouliagos, C. and Jarrell, S.B. (2008). Meta-regression analysis as the socio-economics of economics research. The Journal of Socio-Economics 37(1): 276-292. Stefani, G., Romano, D. and Cavicchi, A. (2006). Consumer expectations, liking and will- ingness to pay for specialty foods: Do sensory characteristics tell the whole story? Food Quality and Preference 17(1-2): 53-62. 154 Fabio Gaetano Santeramo et al. Sutanonpaiboon, J. and Atkin, T. (2012). Using region to market wine to international consumers. Journal of Food Products Marketing 18(1): 1-18. van der Lans, I.A., van Ittersum, K., De Cicco, A. and Loseby, M. (2001). The role of the region of origin and EU certificates of origin in consumer evaluation of food prod- ucts. European Review of Agricultural Economics 28(4): 451-477. van Ittersum, K., Meulenberg, M.T.G., van Trijp, H.C.M. and Candel, M.J.J.M. (2007). Consumers’ appreciation of regional certification labels: A pan-european study. Journal of Agricultural Economics 58(1): 1-23. Verbeke, W., Pieniak, Z., Guerrero, L. and Hersleth, M. (2012). Consumers’ awareness and attitudinal determinants of European Union quality label use on traditional foods. Bio-based and Applied Economics 1(2): 213-229. Viscecchia, R., Nocella, G., De Devitiis, B., Bimbo, F., Carlucci, D., Seccia, A. and Nar- done, G. (2019). Consumers’ Trade-Off between Nutrition and Health Claims under Regulation 1924/2006: Insights from a Choice Experiment Analysis. Nutrients 11(12): 2881. Economics of culture and food in evolving agri-food systems and rural areas Severino Romano1, Francesco Vanni2, Davide Viaggi3 On the relevance of the Region-Of-Origin in consumers studies Fabio Gaetano Santeramo1, Emilia Lamonaca1,*, Domenico Carlucci2, Biagia De Devitiis1, Antonio Seccia1, Rosaria Viscecchia1, Gianluca Nardone1 Benefits for the local society attached to rural landscape: An analysis of residents’ perception of ecosystem services Stefano Targetti1, Meri Raggi2, Davide Viaggi1 Assessing preferences for rural landscapes: An attribute based choice modelling approach Cathal O’Donoghue1, Stephen Hynes1, Paul Kilgarriff2, Mary Ryan3, Andreas Tsakiridis3 Exploring governance mechanisms, collaborative processes and main challenges in short food supply chains: the case of Turkey Yaprak Kurtsal, Emel Karakaya Ayalp, Davide Viaggi