Step interventions and market integration: Tests in the US, UK, and Australian property markets

Market integration implies the existence of some long-run equilibrium relationship between markets such that movements in one market are transmitted to movements in another. It is an interesting observation of much of the literature regarding a possible relationship between real estate and stock markets that there is relatively scant attention given to the possible existence of structural breaks and the impact that such breaks may have on tests for market integration. Other research has shown that failure to take into account structural breaks in various macroeconomic data series may have yielded misleading results on cointegration (in particular, unit root tests on individual series). In this article we examine the issue of whether the stock market and real estate markets are stationary or nonstationary in the presence of structural breaks. We adopt the techniques of Perron (1989), Zivot and Andrews (1992), and Perron and Vogelsang (1992). Each of these tests is based on different assumptions and therefore may yield differing results. In general, the results do not support cointegration of domestic property and equity markets or cointegration of markets internationally.