Covariance Risk and the Ripple Effect in the UK Regional Housing Market

This study combines two increasingly popular areas of the housing literature, by incorporating a measure of the ripple effect into a model of house price volatility. Using UK data for the English regions and Wales from 1995 to 2016, a multivariate GARCH model is initially used to produce time-varying covariances between house prices in London and other regions. These covariances are then incorporated into an EGARCH model of house price volatility showing that this covariance term is highly significant and positively signed in all regions, such that an increase in the covariance with London has a positive effect on a region’s house prices. However the GARCH term in the mean equation produces a negative risk/return relationship across regional housing markets, although it is not robust to different specifications. This suggests that in the UK, when treating regional housing market risk, policies need to include the relationship of the region’s house prices with those of London. Similar considerations could also apply in other countries exhibiting ripple effects in their housing markets.