The role of tourists for pastures resilience and agriculture sustainability in the Alps: a multivariate analysis approach

Traditional extensive agriculture is the main factor of landscape management in mountain: a large part of the Alps is modelled by agriculture. According to Soliva et al. (2008), low intensive farming systems generate extensive green areas, such as meadows and pastures. These systems substantially comprise much of the mountain landscape mosaic, the so-called semi-natural environment, reflecting their anthropogenic nature. Interpreted as a multifunctional activity, including landscape modelling and maintenance, agriculture generates the conditions under which mountain pasture landscape is an economic resource potentially exploitable by tourism. Thus, tourism has acquired a central position in the debate on the future of rural, upland and mountain economies across Europe, especially in the light of both the increasing recognition of these services and functions (Lee et al., 2010) and the demand for the natural environment and landscape (Castellani & Sala, 2009). Some studies concerning mountain rural tourist profiles have been developed during the recent years (Bernuès et al. 2015, Devesa et al. 2010, Mazzocchi & Sali, 2016). The innovation of this work is a mixed approach, based on a multivariate analysis using Multiple Correspondence Analysis (MCA) and a Hierarchical Cluster Analysis (HCA) to characterize the profiles of potential tourists of mountain pastures. Among the characteristics of tourists, our model includes their WTP for pastures resilience. Since the analysis outlines the profile of tourists, it may contribute to a marketing project aimed to communicate and advertise the grazing system and landscape, taking advantage of the demand for tourism in the most efficient way. Moreover, useful guidelines for public policies can be drawn based on the results to invest public resources in the aspects of the good best evaluated by tourists. The analysis has been performed with direct interviews to tourists, conducted in some Italian municipalities of the Seriana, Brembana and Scalve Valley, located within the Regional Park of Orobie Bergamasche Alps, in the northern area of the Province of Bergamo, in Lombardy Region. The survey has been realized from July to December 2015, and 429 interviews have been collected. The structure of the questionnaire is based on closed questions to standardize responses and facilitate categorization. In the first part of the questionnaire subjects were asked to express their opinion (on a scale from 1 to 5) concerning the importance of the environment, territory, accessibility and activity of pastures. In the second part of the questionnaire, willingness to pay was made explicit using the double bounded method. The third part of the questionnaire addressed the personal characteristics of the respondents. The methodological approach presents three steps: the WTP assessment, the MCA improvement, and the HCA. In the present study, a Contingent Valuation Method (CVM) has been used. To assess WTP, the information was directly elicited from individual i after applying a contingent valuation questionnaire using the dichotomous choice model, generating a dichotomous answer (yi= 0 if the individual answers no and yi= 1 if the answer is yes) to a question about paying a previously determined amount si, which randomly varies across individuals (Lopez, 2012). According to Mazzocchi & Sali (2016), we used a double bounded model to introduce the potential identification of two limits, lower and upper, in which the individual WTP is located. The second step of the methodology involves a multivariate analysis, given the considerable number of categorical compared with quantitative variable. Because MCA transforms categorical variables into continuous variables and enables the characterization of groups of individuals based on categories using a smaller number of variables, while retaining as much information as possible, it is widespread practice to use this statistical technique in combination with cluster analysis (Ruggeri et al., 2017). Cluster analysis following a MCA is often used to classify individuals into homogeneous groups; in this case, the clusters are derived from the MCA dimension object scores, which are based on the quantification of the qualitative variables that define the individual profile. Concerning the results, we have obtained three clusters. The first cluster includes respondents with a strong preference for the agricultural aspects of pastures and upland farms. Individuals in this group opt for a pasture increase and a subsequent decrease of the woods, only associating these areas with trails and preferring a management model of upland farms based on agriculture rather than on tourism. Furthermore, this cluster is the only group in which respondents have a high WTP value. This feature delineates a profile highlighting individuals who are sensitive to the issues addressed by the survey, i.e., the survival of the high-altitude pastures and farms since this group combines a high WTP value with the agricultural orientation of the upland farms. However, this group also emphasizes the preference for an infrastructural network based on trails, suggesting the importance of infrastructure in mountain communities. Thus, this group represents the “agro-tourists” cluster. The second cluster includes respondents who envision the best management model of upland farms and pastures for tourism, providing paved roads for pastures to improve accessibility, with no interest in using these lands for grazing; these individuals prefer the reduction of pastures in favour of woods. These respondents are likely more interested in tourist environment activities, such as open-air activities and sports, but not necessarily in those associated with agriculture. Indeed, respondents belonging to this second cluster have a low WTP value and understand the mountain grazing abandonment problem, but their solution for the revitalization of pastures is strictly associated with classic typologies of tourism rather than an agro-touristic model. These individuals likely agree with the total reconversion of some of the pasture structures in touristic areas, where possible, and are not interested in an agricultural approach. The education level among the members of cluster 2 is generally high. Cluster 2 represents the “tourism oriented” cluster. The third cluster includes respondents who are satisfied with the current management of pastures and upland farms, and individuals who do not consider infrastructure to be indispensable for the municipalities of the valley. These individuals typically have low education levels. In addition, this group is not willing to pay for the maintenance of pastures and upland farms, associating these areas only with pedestrian paths. Cluster 3 represents the “traditionalist tourists “. In conclusion, this work suggests a different sensitivity of tourists with respect to the conservation and exploitation of pastureland for tourism purposes, leading to several management methods. Future researches could better explore the differences among the WTP for pasture maintenance, for example also considering different sample and areas.