Abstract:
Recent developments on German real estate markets show a striking increase of rents, especially in larger towns. This development is, however, not homogeneous: The market dynamics vary between different parts of cities and prices develop highly heterogeneously. Therefore, small-scale results are of interest. The German Mikrocensus, Germany's largest household survey, routinely provides information on housing, including rental prices. Since 2018 results are geo-coded, setting the prerequisite for obtaining results at a local level. This specifically strong regional disaggregation obviously results in small sample sizes for the entities of interest. While standard design-based estimators under these condition result in large standard errors, small are estimation techniques may be a solution to nevertheless obtain reliable estimates. Therefore, the present paper explores the opportunity of obtaining very small-scale estimates of average rental prices based on the Mikrozensus employing small area estimation models. The study focusses on the City of Cologne, which provides a broad range of indicators that can be employed as auxiliary information in the models.