In this paper we examine house price synchronization in 15 global cities using real house price data from 1995:Q1-2020:Q2. We find that although there is evidence for bilateral positive phase synchronization, there is no evidence for an integrated global housing market for our sample of cities. Using a hierarchical clustering approach, we identify three clusters of cities with similar housing price cycles that are not solely determined by geographic proximity. We interpret this finding as suggestive of a rather segmented housing market for the global cities in our sample. Using a dynamic factor model with time-varying stochastic volatility we decompose a city's real housing price growth into a global component, a clusterbased component, and an idiosyncratic component. For most cities in our sample, the global component plays a minor role, whereas the cluster-based factor explains a large fraction of the observed variation in real house price growth with its contribution peaking during the Great Recession of 2007-09.