The effect of monetary policy on housing: a factor-augmented vector autoregression (FAVAR) approach

Abstract

This study examines the link between monetary policy and the housing market. The analysis is conducted using impulse response functions derived from a factor-augmented vector autoregression (FAVAR) model. The FAVAR methodology as developed by Bernanke et al. (2005) avoids the degrees of freedom problem present in standard vector autoregression (VARs) models. The estimations are conducted using 120 macroeconomic time series in monthly frequency for the period January 1959 to August 2001. Results indicate that housing starts respond negatively to monetary policy shocks. This result is consistent across regions in the United States. In the case of housing permits and mobile home shipments, the response to a monetary policy shock is positive at first, but becomes negative after a few periods.

Acknowledgements

The author wishes to thank Isabel Ruiz and Mark Wheeler for helpful comments and suggestions.

Notes

¹The conference board is a nonprofit organization that distributes information about the economy. It is a widely quoted private source of business intelligence.

²Stock and Watson (2002, 2005) suggest some guidelines for determining the number of factors to include in these models. Their arguments are based on the information criteria created by Bai and Ng (2002). Bernanke et al. (2005) argue that the information criteria created by Bai and Ng (2002) does not address the issue of how many factors should enter F_t . They propose a comparison of the results using different numbers of factors.