Real estate prices are central to a range of themes that are, e.g., relevant for monetary policy, community development, environmental valuation, and economic planning more generally. This paper developes a real estate index based on apartment offer prices on the post code level for Germany, taking into accout apartment heterogeneity and small sample sizes within regional areas as well as spatial and temporal dependencies. In a first step, a hedonic price function is estimated. In a second step, the residuals calculated from the hedonic function are used as direct estimates in a small area estimation (SAE). This technique is designed to yield estimates with a smaller variance in the context of small samples. The results show similarities between the estimates obtained from the residuals and SAE estimates. But the SAE models show non-negligible gains in accuracy for the coefficient of variance, i.e. the estimates are stabilized.