Modeling multivariate tourism expenditure using vine copula: empirical findings from of Fribourg-Switzerland.

The measurement of economic contribution of tourism in a given geographical area needs, besides non-monetary measures such frequentation ones, a monetary one namely tourism expenditure. In addition to these macroeconomic related purposes, the whole statistical tourism measures shed light on tourist behavior, which are of great interest for marketing analysis. This study proposes to model the joint distribution of tourism expenditure disaggregated by category simultaneously with a set of covariates associated with the tourist's trip (including destination, days of stay, season of the year, among others). The dependence structure, among all variables, is established by a family of pairwise copulas (regular vines), which allows fitting a high-dimensional multivariate statistical model. As an application, the model is calibrated with a database for the Fribourg region (Switzerland). The good fit of the model to the data is observed both in the marginal distributions and in the dependence structure. The parameters representing dependence of the vine copula show strong relationship between the different categories of spending, both for tourists and excursionists. The empirical results show that the key variables to understand the associations between them are the destination in the region and the place of origin of the visitor. In general terms, the empirical results show that a visitor that expends more on one item is likely to expend more on a complementary items of spending. [ABSTRACT FROM AUTHOR]