Predicting vacancy from administrative data
Until recently, very little was known about the extent of vacant housing in the Brussels Capital Region (BCR). At the request of Bruxelles Logement, the BSI, in collaboration with the Brussels Informatie-, Documentatie- en Onderzoekscentrum (BRIO, VUB) and the Institut de Gestion de l’Environnement et d’Aménagement du Territoire (IGEAT, ULB), looked into the possibility of predicting the probability of a specific property in the BCR being unoccupied by linking existing administrative data. Specifically, the linking of four databases (i.e. the land register, the national register, the Crossroads Bank for Enterprises in Belgium and data on low water consumption (Vivaqua)) combined with site checks by the administration on a random sample of 5 000 buildings provided the basis for a usable model. In this respect, an empty housing unit can be understood as a housing unit with no domiciles, low water consumption, where a business may be located…
Need for unique property identifier
Linking the different datasets was no easy task. Indeed, in the absence of a unique housing identifier, each data file containing information on housing units uses its own identifier, in practice the address. However, the way in which these addresses appear in the files varies greatly, with the result that linking the various files was an enormous task.
Vacancy detection tool
A useful by-product of successfully estimating a number of vacancy models, is the development of a vacancy detection tool that allows to predict the risk of vacancy. This allows the administration involved to be much more targeted in detecting vacant properties.
Estimate number of vacant properties
Finally, based on this study, estimates of the number of vacant properties in the BCR can also be made. The analyses do suggest that what is sometimes interpreted as a vacant property is, in fact, a used property, but one without domiciles (think tourist accommodation) or one used for non-residential purposes. Encouraging the ‘right’ use of housing therefore seems equally important. Taking this into account, the model suggests that the BCR has some 700-9 000 dwellings that are ‘vacant’ without a valid reason (‘sanctionable vacancy’). This study is, to our knowledge, the first to estimate the level of vacancy in the Brussels Capital Region based on a random sample.