This paper is mainly concerned with the analysis of regional house price cycles. Based on a newly available data set consisting of the 40 largest U.S. Metropolitan Statistical Areas (MSAs), we introduce a wavelet transform based metric to study the housing cycle synchronization across MSAs. We derive several conclusions: (i) We show that regional housing cycle dissimilarities are significantly and strongly connected to geography. (ii) We show that U.S. regional housing cycles are considerably shorter compared to business cycles. (iii) By employing statistical methods, we also show that regional housing prices significantly converge after the bursting of the bubble only at higher business cycle frequencies, whereas for lower business cycle frequencies this is not the case. This is a notable result, because it directly implies that housing cycles behavior is different even at the business cycle.