Spatial and temporal diffusion of house prices in Sweden

Abstract: In Sweden, both house prices and household debt have more than doubled over the last two decades. This rapid surge has generated concerns about a housing bubble and the economic consequences that may follow, resulting in a considerable body of research analysing the fundamentals of Swedish house prices. To assess how a bubble or a crisis would ripple over the housing market, it is key to also understand the dynamics of house prices. This paper examines the spatial and temporal diffusion of house prices in Sweden. The first aim is to empirically determine if regional house prices respond to a change in Stockholm house prices. The second aim is to establish which measure of proximity that gives the most accurate representation for the cross sectional dependence between regions. The purpose is to assess how a change in Stockholm house prices could ripple over Sweden in the event of a bubble or a crisis. The analysis is performed by estimating two price diffusion models with OLS, using monthly regional real house price data over the period 2005m1 to 2018m12. The results suggest that Stockholm affects house prices in more than half of the other regions contemporaneously, and that regions that are economically or geographically close to Stockholm are more affected than regions far away. Further, the results suggest that economic proximity is better than geographical closeness at explaining the cross sectional dependence between regions, resulting in a more reliable analysis of the spatial and temporal diffusion of house prices. Together these results provide important insights for how a change in Stockholm house prices could ripple over Sweden in the event of a bubble or a crisis.