Abstract

We present the methodological underpinnings and the compilation process of quarterly price indices for the German commercial real estate (CRE) market. We use a dataset maintained by the Association of German Pfandbrief Banks (Verband deutscher Pfandbriefbanken, vdp). It contains CRE transactions of office buildings, retail structures and multi-family dwellings. The information is taken from the financing business of banks. We compile price indices for the three object types separately and as a total. The volume and coverage of the dataset enable us to provide a regional breakdown using a meaningful spatial typology distinguishing between several city groups and urban or rural areas. As CRE is characterised by high heterogeneity and data is sparse, practical index compilation is challenging. We elaborate on our modelling choices as regards stratification and hedonic quality adjustment. We employ filtering techniques in order to separate meaningful price signals from random noise in the data. We further illustrate how volatility and filtering affect the revision policy of the CRE price indices.