Unit root test under the structural change and applications on macroecoomic variables

Time series variables can include structural breaks by some reason which can be wars, peace, change of policy implementations, economic crises. Structural changes that are accuring in the series cause diffucilties of determining their stationrity. If one apply unit root tests without taking notice of these structural changes the results can have errors and these affect the power of the tests. In that case, a time series can be specified to have even though the series do not actually have unit root. For obtainig reliable regression results one need to take structural changes into account.In this study, testing the steadiness of particular Turkish macro economic data under structural break was intended. In this context, whether the series include unit root and the effects of structural breaks in the series' trend function on the unit root process were examined. The sample covered quarterly data for 1987-2009 period. Firstly we analyzed the series by using of standart unit root tests; ADF, PP and KPSS, which do not take precsence of structural change inot account. Then, relevant macro economic variables were analyzed with some particular unit root tests where the structural breaks are not taking into account. In order to analyze the effects of one time break in the series' trend function on the unit root tests, test methods powered by Perron (1989), Zivot and Andrews (1992), Perron (1997) were conducted. Moreover, tests used by Lumsdaine and Papell (1997) and Lee and Strazicich (2003) to test more more than one break in the series' trend function were examined. In this study, observation of GNP, consumption, interest rate, gold prices, ISE 100 index, Money supply (M1 and M2), wholesale price indeks, Costumer Price Index and Exchange rate were applied.Keywords: Time Series, Stationarity, Unit Root, Structural Break