Regional House Prices: An Application of the Swedish Housing Market

A variety of factors explains the differences in the prices of single-family houses. Changes in the gross domestic product over time are one important determinant together with changes in the interest rate. However, the scope of this paper is to explain why house prices differ from region to region. The paper follows the traditional economic theory (see e.g. DiPasquale and Wheaton, 1994) used to explain aggregated housing prices over time. That is, the developments of the house prices are a function of macroeconomic factors such as economic growth and changes in employment and interest rate. We estimate a two-equation model: first, the long-run price equation and, second, an error correction model. In this study we also estimate a spatial autoregressive model. We analyze a panel data set consisting of 2,810 observations. Our main contribution is that we have estimated the speed of adjustment to the long-run equilibrium price for 21 regions instead of at a national level. We observed that the rate of adjustment to equilibrium price the first year varies considerably from region to region. Our conclusion is that the speed of adjustment ranges from 10Ò70 percent (around 28 percent on average) depending on the region. In regions with a low population density, we observe higher price adjustment rates. Furthermore, if the model includes a spatial autoregressive component, the parameters change only modestly compared to an OLS fix effect model. Finally, the out-of-sample prediction indicates that the regional housing price model with individual price adjustments outperforms na‘ve predictions.