ABSTRACT
During the last years, the implementation of large tourism behaviour surveys organized at national level has been generalized in many countries. However, since many of the tourism policies are decided at a local level, regional planners and managers frequently question the usefulness of these national surveys when, in fact, the number of available surveys representing their regions is too small to get a representative view of the tourists’ profiles in their specific destination or business. This work presents and applies small area techniques to overcome the problem of obtaining estimators of certain variables of interest for very specific market segments, taking advantage of the cross-information that defines the segment. To illustrate the technique, the Spanish Household Budget Survey is considered, in particular, two specific categories, related to tourism spending, illustrating the potential of the methodology in travel and tourism planning. The results show the possibility of extracting expenditure information in market segments even in the absence of sample observations.
Acknowledgements
We acknowledge the Agencia Estatal de Investigación (AEI) and the European Regional Development Funds (ERDF) for supporting the project: PID2019-106738GB-I00 /AEI / 10.13039/501100011033 and ECO2017-83255-C3-2-P/AEI / 10.13039/501100011033.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 In 2016, 2017 and 2018 for Spain there are 22,011, 22,043 and 21,395 households in the sample, representing 18,444,023, 18,512,537 and 18,627,677 households in the population, respectively. The chosen areas for the case studies, understandably, have much smaller sample sizes. For the given years, the Balearic Islands have 763, 753 and 633 observations, which represent 446,498, 448,131 and 453,533 households in the population. The city of Madrid has 863, 865 and 796 households in the sample, representing 1,428,950, 1,396,999, 1,388,521 households in the population. And, finally, the non-urban Castilla-León region has 814, 812 and 784 observations, corresponding to the 527,262, 528,581, 539,830 households in the population, for 2016, 2017 and 2018, respectively.
2 The classification is based on the original variable TIPHOGAR5 which was modified accordingly. Original categories 1–4 were merged into category (a), categories 5–7 merged into category (b), category 8 renamed as category (c) and categories 9–12 merged into category (d). Initially there were some households with missing classification of type TIPHOGAR5, but they had information about other classification types, so we were able to classify all households in the above four types.
3 The classification is based on the numerical variable EDADSP (age of the main provider), which was modified accordingly to create 3 age intervals. The cutoff age of 65 is used in the Spanish HBS, meanwhile the cutoff age of 45 commonly used by the National and European statistical agencies (INE, Eurostat).