Panel unit root test based on seemingly unrelated regression models that via residuals augmented with Fourier function

Stability tests are important in the time series and panel data literature. Because econometric analyzes to be made with non-stationary series can cause false regression problems. In this context, the stationarity tests developed with unit root tests have become systematic. Time series can experience ups and downs due to different reasons during long periods. Natural disasters, wars, economic crises, policy changes are some of these reasons. For this reason, the ups and downs coming from the serendipitous are called structural change or fracture. These structural changes mentioned may occur in the average of the series as well as in the trend. To ignore in unit root tests of breaks from the average of the series or its trend may cause to get sloping results. Conventional unit root tests applied to the series containing structural change will lose their validity. This study shows how the SURADF that is one of the second-generation unit root tests proposed by Breuer vd. (2001 2002) unit root test and based on seemingly unrelated regression models (Seemingly Unrelated Regression- SUR), is extended to the structure taking structural changes into consideration. Chang (2012) added fourier functions to take account of structural changes in the SURADF unit root test. In addition to considering the structural changes with fourier functions, the panel unit root test, which gives stronger test results by adding the models to the normal non-scattering information of the errors, has been developed. The process recommended by Enders and Lee (2012) was followed in the model of the deterministic component using Fourier functions that can catch changes correctly, ın cases where the number of structural changes, their location and structure are unknown. Thanks to the Fourier functions, smooth transitional structural changes can be detected more correctly. In addition, the stronger unit root test results were obtained by including to the model information of normal non-distribution of faults in the unit root test study run by Im and Schmidt (2008). In this study, the power of the unit root test It has been increased by including to the SURADF with fourier functional RALS estimators, which are stronger predictors in the case of normal non-distributing faults proposed by Im and Schmidt (2008). Key words: Unit Root, Panel Data, SURADF, Fourier, RALS.