A Closer Look at the U.S. Housing Market: Modeling Relationships among Regions



This article investigates the dynamic interactions among nine U.S. regional housing markets by estimating the multivariate cointegration model using both autoregressive (AR) and moving average (MA) representations over the period from 1975 to 2010. Long-run results indicate that the extent of convergence among the regional housing markets substantially increased over time and more so after the housing bubble burst in the latter part of 2006. Common stochastic trend analysis reveals that the housing regions of New England, Mid-Atlantic and the Pacific were the primary regional drivers that led the regions toward long-run equilibrium during the 1975 to 2006 subperiod. Further analysis indicates that the relationships among the regions cannot be attributed to trends in two important macroeconomic fundamentals: regional per capita income and regional GDP. Finally, short-run analysis reveals substantial lead lag relationships among all the markets.