Regional house price behaviour in the UK: application of a joint testing procedure

Abstract

The notion of a ripple effect in the UK housing market implies stationarity in regional:national house price ratios. In this paper a new means of examining this issue is proposed which involves the joint application of a powerful unit root test and a test of stationarity. In contrast to the previous studies which have failed to detect stationarity using the Dickey–Fuller unit root test, the proposed method uncovers stationarity for a number of regions. Monte Carlo evidence is presented to support and explain the empirical findings obtained. Importantly, it is found that in comparison to previously considered joint testing approaches and individual unit root tests, the proposed method dramatically reduces the frequency with which stationary series are mistakenly concluded by practitioners to possess a unit root.