Hedonische Regression der Wohnungsmietpreise unter Berücksichtigung von Lagevariablen am Beispiel eines Bestands im Ruhrgebiet

In diesem Artikel wird der Einfluss makroökonomischer Lagevariablen auf die Markt- und Kostenmiete in zehn Städten des Ruhrgebiets untersucht. Der verwendete Paneldatensatz umfasst Daten zu 27.336 Wohnungen für die Jahre 2000 bis 2004. Neben einer Reihe von kontrollierenden Wohnungseigenschaften wird der Einfluss der makroökonomischen Variablen Altersstruktur, Arbeitslosenquote, Ausländeranteil und Anteil der sozialversicherungspflichtig Beschäftigten auf die entsprechenden Mietpreise getestet. Untersuchungsebene sind die statistischen Bezirke der jeweiligen Städte. Darüber hinaus werden die Wohnungen im südlichen Ruhrgebiet mit einer zusätzlichen Indikatorvariablen erfasst, um eventuelle Preisdifferenzen zum nördlichen Ruhrgebiet aufzuzeigen. Unsere Regressionsergebnisse legen einen eher homogenen Wohnungsmarkt im Ruhrgebiet nahe. Bestätigt wird hierbei die Annahme, dass Mietpreise vor allem durch die Wohnungseigenschaften und weniger durch Lagevariablen beeinflusst werden. Ersichtlich wird zudem, dass die Mietpreise für Sozialwohnungen (Kostenmiete) noch weniger von den untersuchten Variablen abhängen und somit losgelöst von der Preisbildung auf dem Markt entstehen. This study seeks to examine how various characteristics of rental housing influence rent prices. These characteristics are divided into both dwelling-specific and macroeconomic variables so that their influence on both open-market rent and the legally-specified rent covering total investor‘s expenses (the so-called „Kostenmiete“, the rent chargeable for social housing in Germany) in ten cities in the Ruhrregion could be analysed using a multivariate regression (hedonic regression model). In addition to a series of controlling factors for the residential properties (size and age of the house, number of rooms and an indicator variable for refurbishment), we have tested the influence of macroeconomic variables: age structure of the inhabitants; unemployment rate; proportion of immigrants; and proportion of the working population (denoted with statutory social insurance coverage) on the rent level. Furthermore, dwellings located south of the Hellweg - a historically important supraregional route connecting Western with Middle and Eastern Germany that crosses the Ruhr region - have been marked with a dummy variable to test the presumed higher quality of living in this Southern part of the region. The main findings of the study are that all building related variables, here the size and the age of the house, the number of rooms and an indicator variable for refurbishment, have a significant influence on the rent level. The direction of this influence is in accord with our expectations and the literature. The rent price is increasing with size of the flat and for renovated houses and it is decreasing with a ceteris paribus higher number of rooms and the age of the house. Furthermore, it can be stated that the „Kostenmiete“ as the rent for social housing is less scattered than are open-market rents. This may be explained by the fact that the market mechanism for rented accommodation is being neutralised to some degree. The positive and negative influence of the houses‘ characteristics is confined due to the legal restrictions following the „Kostenmiete“-concept. We observe two different rent-price regimes for the open-market and the socially rented houses. In contrast to the variables for the building characteristics, the accommodation-related variables for age distribution and high proportion of immigrants in the population do not have a significant influence on open-market rents: only the indicator variables for a high proportion of the working population and for location south of the Hellweg region have proved to be significant. The insignificance of the socioeconomic variables concerning inhabitants may be due to the sample of a housing company and the rather aggregated data on statistical districts. However, an alternative explanation is that these factors do not influence rents within this region and are not important enough to influence the rent prices for the flats in the sample.