There is scant empirical support in the literature for the Fisher effect in the long run, though it is often assumed in theoretical models. We argue that a break in the cointegrating relation introduces a spurious unit root that leads to a rejection of cointegration. We applied new break tests and tested for nonlinearity in the cointegrating relation with post-war data for 15 countries. Our empirical results support cointegration, after accounting for breaks, and a linear Fisher relation in the long run. This is in contrast to several recent studies that found no support for linear cointegration.