Some recent time series studies testing the stationarity of real exchange rates (RERs) produce conflicting results. Using nonlinear unit root tests and recursive analysis, this paper tests whether the evidence on the stationarity of RERs is sensitive to different numeraire currencies, different sample periods covering regional and global crises, and the inclusion of countries with different levels of economic or regional integration. The results indicate that evidence for a stationary RER could be substantially sensitive to sample period changes, but not so for the currencies of the countries involved in forming the euro area. We also find that financial crises have a notable impact on testing the stationarity of RERs, depending on the numeraire currency used. We discuss the policy implications of the findings.

We use rolling cointegration to measure the convergence of base money, M2, the CPI and industrial output between Germany and France and recent EU members and some transition countries that are now joining the EU. Countries that joined the EU previously exhibit time-varying cointegration with the core countries over the 1980–2000 sample period. Cointegration for the transition economies was comparable for M2 and prices, but not for monetary policy and industrial output. Thus a peg to the Euro soon after accession is feasible for the East European countries, but the benefits of joining the Euro zone are as yet limited.