It is challenging to estimate granular-level indices of infrequently traded assets because data can be extremely scarce—the degree of freedom can be near zero if the estimation only uses local properties. This article applies a parameter-reduction approach to U.S. commercial real estate data, and estimates metro-level total return indices by property types from 1997 to 2014. This article further evaluates the economic merits of the indices. Test results suggest that the estimated metro indices have significant explanatory power, both in- and out-of-sample, for local property returns.
All Science Journal Classification (ASJC) codes
- Economics and Econometrics