| Download Document | |
| Document Type: | General |
| Publish Date: | September 2010 |
| Primary Author: | Patrick J. Lawler |
| Prepared By: | Andrew Leventis |
| Edited By: | Suneela Farooqi |
| Published By: | Patrick J. Lawler |
This paper describes a new methodology for estimating U.S. average and median house prices. The approach, which relies heavily on repeat-transactions house price indexes, attempts to construct statistics that are less vulnerable to certain types of distortions than existing metrics. Problems arising from changes in the geographic composition of the underlying data sample are particularly challenging to overcome. For example, the calculation of average home values for a given state can be problematic because within-state transaction volumes may disproportionately represent certain areas in a given period. Calculating reliable “median” or “mean” home values for any geographic area is difficult, particularly if the area is large and includes homes in heterogeneous neighborhoods. Assuming the goal is to produce summary home values for the underlying housing stock, and not the limited exercise of producing summary statistics for transacting properties, a number of problems can arise.