Testing for structural breaks in the Korean economy 1980-2005: an application of the innovational outlier and additive outlier models

This paper employs quarterly time series data to endogenously determine the timing of structural breaks for various macroeconomic variables in the Korean economy. The Innovational Outlier (IO) as well as Additive Outlier models (Perron, 1997) are then used to test for non-stationarity of the Korean macroeconomic data. After accounting for the single most significant structural break the results from the (AO) model clearly indicate that the null o f a unit root cannot be rejected for all of the series under investigation. This finding is consistent with our finding based on the conventional unit root test. However, by applying the IO procedure in the presence of a structural break we find the interesting result that two of the variables under investigation become trend stationary (have no unit roots). The timing of structural breaks for key macroeconomic data under the IO and AO approaches appear to be quite different. Using the IO approach seven of the ten macroeconomic variables focused upon have important structural breaks corresponding with the timing of the Asian financial crisis of 1997. On the other hand, using the AO approach, only one of the ten variables appears to have a structural break related to the Asian financial crisis, while the remaining nine variables have quite diverse structural breaks that depend on key policy changes or other factors contributing to economic turbulence.