Unit root tests and cointegration test under the structural breaks:Money demand stability

If there is a break in time series used and unit root is done without considering this break, then it is seen that there is strong bias of the series not to be stationary. In this study, Perron (1989), Zivot and Andrews (1992) and Peron (1997), which are unit roots that are developed where there are structural breaks, Gregory - Hansen (1996) methods that are cointegration tests are examined. To test these methods, for 1987-2007 era with three-month datas, money demand stability model is used for Turkey. In a limited sense money demand stability is tried to be modelled by Gross Domestic Product (GSYİH) as income variable and three-month deposit rate. Although the financial crisises of 1994, 2000 and 2001 in Turkey caused structural breaks, Zivot and Andrews (1992) and Perron (1997) unit root results, did not change the result that series include unit root. The relationships between variables are analyzed by Gregory and Hansen (1996) cointegration test. According to Gregory and Hansen (1996) results, it is determined that there is no meaningful relation of cointegration between money demand under structural break, Gross Domestic Product (GSYİH) and three-month deposit rate.