"Gimme Shelter" - A Long- and Short-Run Analysis of Swedish Regional House Prices

The aim of this thesis is to examine the long-run relationship between regional house prices, how the house price model behaves in the short run and if aggregated national house price models may be misleading. To investigate whether house prices are related in the long run we use cointegration tests to look for common trends. The short-run behaviour of the model is estimated by an error correction model. To test the hypothesis that aggregated models are misleading, we compare the regional models and also use the same procedure to estimate the house price model with panel data. The results show that there are several long-run relationships between house prices of different regions. In the short run, CPI seems to have an important impact on house price changes in all regions whereas GDP and unemployment mainly influences house price movements in the metropolitan areas and the southern half of Sweden. The results from the panel regression lead to similar conclusions regarding the independent variables. Although the results are similar regarding the statistical significance of the variables, we cannot say anything regarding differences in economic significance between regions when using panel data and an aggregated model may thus be misleading