Abstract
This research employs a filtered and unfiltered value at risk (VaR) model to evaluate the downside risk in housing markets in the United States and the United Kingdom. Empirical results show that the filtered general Pareto distribution (GPD) can correctly capture the downside risks in both housing markets. As the actual return distribution in the U.S. housing market is non-normal, the model of normality assumption underestimates extreme risk in this market. Finally, the value at risk (VaR) of filtered models can mirror the dramatic change in downside risk in the housing market. Hence, VaR can be used by mortgage banks to monitor foreclosure risk to prevent the unfavorable impact of systematic risk.