2009
Abstract:
The U.S. housing price indexes are subject to measurement problems
that severely impair their ability to capture the true risk. In this paper, we
seek alternative methodology of utilizing latent-variable statistical methods
and provide new insight to understand housing market dynamics. Housing
prices are assumed to respond to external forces as proxies by way of a set
of macroeconomic variables and Önancial indexes. Latent variable models
allow us to extract interpretable common information about unobserved real
estate returns.
This methodology of this paper is based on the framework in Bai & Ngís
papers (2002 &2006). We applies a pure statistical approach to extract the
latent factors and, more importantly, examines whether the observed macro
factors are exact factors according to the information criteria proposed by
Bai &Ng (2006). For robustness, we test both OFHEO repeated sales index
and MRAC median home price index. We found geographical pattern of
factor loadings for housing price appreciation at MSA level. The results
indicate that income, consumption and GDP is a comparatively accurate
factor.
