Multi-scale causality and extreme tail inter-dependence among housing prices

Abstract

This study explores multi-scale causality and extreme tail dependence structures among housing prices in four cities: Seoul, Hong Kong, Tokyo, and New York. We apply two different and unique approaches in our analysis of monthly housing price data: (i) the frequency domain Granger casualty test and (ii) the non-parametric copula test. Employing the frequency domain casualty test, we find both bi-directional and uni-directional causalities at different frequency bands. Additionally, the nonlinear copula estimates indicate asymmetric tail dependence for housing price pairs in all four cities. Finally, the Hong Kong housing market has a greater effect on the Seoul and Tokyo housing markets than does the New York housing market.