An Answer To The "Fisher Effect Puzzle": Evidence From The Nonlinear Co-integration Tests

Fisher effect hypothesis has been always a hot-debated topic. This hypothesis statesthat the nominal interest rates will follow the change of expected inflation in a one-to-oneway, which means the real interest rates will never change during the whole process. Thatis to say, if this fisher effect holds water, monetary policies will be of neutrality.Previousstudies assuming that the fisher relationship is linear fail to find hard evidence for thishypothesis. This is so-called the “fisher effect puzzle”. Recently, many new researchesargue that the nominal interest rates and inflation may have a nonlinear long-runrelationship which may help us to solve this puzzle. Therefore, in this paper, we will applya nonlinear co-integration test to the fisher relationship on the basis of these new nonlineartheories.In the first part of this paper, we will detailedly present the significance of such topicand also an introduction for previous studies of fisher effect from all over the world.After that, in the following part, we will show the development of nonlinearco-integration tests where I roughly divide these tests into three categories:1,theresidual-based nonlinear co-integration tests.2,the ECM-based nonlinear co-integrationtests.3,the tests for co-integration with nonlinear co-integration vector.Finally, the empirical research of this paper investigates seven OECD countries,using a nonlinear STAR error correction model and test. And we eventually find obviousnonlinear fisher effect in these countries.