Modeling the United States Housing Sector

Abstract

Abstract New residential construction and improvement of existing housing account for most of the solid wood products consumption in the United States. It is important for forest policy analysts and those interested in wood products markets to understand the implications of changes in macroeconomic variables on housing markets and, hence, on wood products markets and forests. In past studies of U.S. housing markets used for long-run forest policy analysis, the time series properties of the macroeconomic data used to estimate the models were ignored. To the extent that estimation results are sensitive to those properties, the resulting model forecasts may be misleading. This study reports a structural model of the housing sector in the United States estimated using five alternative specifications to account for the time series properties of the data. Estimation results were compared, and model forecast performance was evaluated. Several stable estimation results emerged. However, further investigation is indicated for some potentially important relationships that were sensitive to model specification. FOR. SCI. 47(3):371–389.