A new nonlinear unit root test proposal based on the interpretation of the MTAR model with fourier functions

When the time series studies are examined, it is seen that there are many linear and nonlinear unit root tests. These unit root tests differ from each other depending on the model structure and structural break used in the testing process, or one may be stronger than the other. The aim of this study is to introduce a new nonlinear unit root test to the literature. In this direction, in the proposed new unit root test process, nonlinearity is modeled using the momentum threshold autoregressive model (MTAR) proposed by Enders and Granger (1998). Structural breaks are investigated with Fourier functions. In this study, a procedure for Fourier Enders and Granger unit root testing is proposed. Critical values are produced using Monte Carlo simulations. In finite examples, size and power properties are examined. According to the empirical results obtained, it can be concluded that the proposed Fourier MTAR unit root test has stronger results than the study of Enders and Granger (1998), when the model with a trend structure is used.