A new nonlinear cointegration test proposal based on fourier functions

The aim of this study is to introduce a new cointegration test to the literature. The new cointegration test developed within the scope of the study is a nonlinear test based on Fourier functions. With this test, possible structural changes are modeled with the help of Fourier functions and nonlinearity is modeled with the Exponential Soft Transition Autoregressive (ESTAR) process in determining the long-term relationships between variables. It has been determined by both Monte Carlo simulations and empirical application that the developed test is a stronger test than alternative tests. In the third part, where nonlinear cointegration tests based on Fourier functions, which form the basis of the study, were developed, 4 different test statistics were developed and critical values corresponding to non-standard distributions of test statistics were produced. Size analyzes and power analyzes of 4 different test statistics were also performed, and it was determined that the new tests were generally more powerful than the alternative tests. It has also been shown that the new tests do not have any size distortion. In the empirical application part of the study, the sustainability of the current account deficit in OECD countries is investigated both with the cointegration tests developed in the study and with alternative tests. The empirical results show that the new cointegration tests have stronger results empirically than the alternatives.