German Real Estate Prices and Rents Across Space and Time

This project will first document the evolution of house prices and rents in Germany in the period 2007-2017 at a quarterly frequency and at a low level of geographical aggregation (municipalities and zip codes), using a large sample of real estate advertisements on the internet platform ImmobilienScout24 and controlling for several quality characteristics of the posted objects. In addition, we are going to document indicators of housing supply (number of postings relative to the stock, and the age structure of the posted units) and housing demand (as measured by contacts and hits for the postings) in each local residential market. Local housing market tightness (separately for sales and for rents) can further be proxied by the observed duration of the postings. We are going to access the data through a scientific use file, available via the RWI Essen. At the same time, we can access the full dataset set in collaboration with Philipp Marek (Financial Stability Division of the Deutsche Bundesbank) who is already working with these data in order to develop high-frequency hedonic price and rent indices for local (district-level) housing markets in Germany.

In a second step, we will investigate the role of house price and rent movements for the evolution of observed patterns of residential segregation across different residential markets and homeownerhsip patterns within residential markets. To do so, we are going to augment our data with regional income and housing variables from several other data sources (such as IAB-SIAB and the German Microzensus).

Then, we are going to construct a spatial equilibrium model with frictional housing markets in which ex-ante heterogeneous households in terms of productive ability choose a residential location and a homeownership status based on the degree of local amenities, opportunities of production and housing market conditions. The evolution of house prices, rents, residential segregation and homeownership across time and space is an endogenous outcome of the model. To match the observed data patterns, we are going to estimate the model’s parameters for preferences, local amenities, local housing market frictions and production opportunities. The model will help us to understand what generates the observed patterns of residential segregation over time and across residential markets. We also examine what are the forces behind the differential changes in homeownership across time and space.