Spatial-temporal asymmetry, shock and memory: housing transaction prices in Sweden
Time: Thu 2021-09-23 12.00 - 13.30
Lecturer: Aleksandar Petreski (Jönköping University)
abstract: This research analyzes volatility clustering of housing prices in time and space for different housing segments. By using transactions data instead of indices and averaged data, the problem of smoothing and reducing the total variance in the data is avoided. Data on infrequent non-repetitive housing transactions is modeled with a spatial-temporal form of ARCH/GARCH. Spatial-temporal weights allow past errors and past conditional variances in neighboring locations to have differential impact on current variance estimated at separate locations. Estimated parameters based on municipal housing transactions from the town of Jönköping in Sweden, from 2015 through 2019, show positive memory effect along with positive shock and confirm the assumption of spatial-temporal persistence in the variance. The market for houses has different spatial-temporal dynamics compared with the market for apartments, showing lower memory (persistence) of the past close-by shocks. Except for the general apartments market, findings show that memory of the price shocks is not greatly dependent on the geographical distance. Only a spatial asymmetry of price shocks is detected.