As the long-term elasticities and the short-term dynamics of housing prices are expected to exhibit substantial regional variation, the conventional panel data models that assume similar dynamics across regions generally are likely to be too restrictive. The aim of this study is to add to the scarce literature that utilizes panel modeling techniques which allow for regional heterogeneity in the housing price dynamics.

We investigate the extent of regional differences in both the long- and short-term dynamics of housing prices across the 50 largest U.S. metropolitan statistical areas (MSAs) using quarterly data for the period 1980-2013. We apply the Common Correlated Effects (CCE) estimators of Pesaran (2006) and Chudik and Pesaran (2015) to avoid another common complication in the extant panel analyses, i.e., the potential bias caused by spatial dependence of housing market variables. In addition to eliminating strong and weak forms of cross-section dependence in large panels, these CCE estimators allow for heterogeneity in the dynamics across MSAs.

We document considerable variation in the long-term elasticity of housing prices with respect to income and in the short-term dynamics across MSAs. The results also provide evidence of cointegration between regional housing prices and aggregate income. The findings have predictability and policy implications.