Content area
Abstract
During the post-World War II period, the residential construction industry experienced a series of fluctuations in building activity. This dissertation provides an empirical analysis of single-family residential construction during the 1971-1990 period, explains the underlying causes of the macroeconomic variables on the fluctuations in new single-family housing construction cycles from the view point of builders, and develops a multivariate time series model that has a better predictive ability in housing starts.
The multivariate time series model which serves as the basis of the study analyzes the new single-family housing markets. The model contains variables for both the demand for and supply of single-family housing construction. The effects of changes in interest rates and credit availability are differentiated. Also, the model contains a variable to show the impact of changes in the builders' expectations for future sales. Furthermore, the Simultaneous Transfer Function (STF) model is employed to improve the lag structure with respect to the cross-time relationships among variables.
The estimation results attest to the validity of the model and clarify a number of basic issues concerning the fluctuation of single-family housing construction cycles. First, the results indicate that while mortgage interest rate, past period sales, inventory, and expectations for future sales are the dominant factors behind the fluctuations, other factors also contribute to the housing sector's instability. Second, the results provide that credit availability has an impact on the construction by influencing the effect of mortgage interest rates. Finally, the results provide that the expectations for future sales play an important role on directly influencing the behavior of the speculative single-family residential builders.