This paper employs newly developed techniques of nonlinear cointegration analysis to study international real estate market integration and price discovery. The countries considered in this study are UK, USA, Netherlands and France. Direct (NCREIF, IPD) and indirect (EPRA) real estate price indexes are used in both linear and nonlinear cointegration tests on bivariate and a variety of multivariate models. The objective is two-fold: a) to examine whether there is price discovery between the direct and indirect real estate market b) to determine the degree of integration between international real estate markets. The models will be evaluated within a portfolio investment framework. Much more evidence of market integration emerges from nonlinear cointegration analysis than linear analysis. It appears, therefore, that many of the conclusions reached in prior work that have used traditional methodologies need to be reconsidered. The degree of market integration has significant implications for portfolio investment where managers seek to develop well diversified portfolios.