Please use this identifier to cite or link to this item:
Kholodilin, Konstantin A.
Mense, Andreas
Year of Publication: 
Series/Report no.: 
DIW Discussion Papers 1207
In this paper, we make multi-step forecasts of the monthly growth rates of the prices and rents for flats in 26 largest German cities. Given the small time dimension, the forecasts are done in a panel-data format. In addition, we use panel models that account for spatial dependence between the growth rates of housing prices and rents. Using a quasi out-of-sample forecasting exercise, we find that both pooling and accounting for spatial effects helps to substantially improve the forecast performance compared to the benchmark models estimated for each of the cities separately. In addition, a true out-of-sample forecasting of the growth rates of flats' prices and rents for the next six months is done. It shows that in most cities both prices and rents for flats are going to increase. In some cities, the average monthly growth rate even exceeds 1%, which is a very strong increase compared to the overall price level increase of about 2% per year.
Housing prices
housing rents
dynamic panel model
spatial autocorrelation
German cities
Document Type: 
Working Paper

Files in This Item:
635.29 kB

Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.