A Comparative Study of the KPSS and ADF Tests in terms of Size and Power

This thesis investigates through simulation why tests of unit root and stationarity occasionally result in different conclusions. The thesis focusses on the KPSS test and the ADF test and both review cases with and without a trend. The goal is to bring additional knowledge of whether one of the tests are more reliable in terms of size and power and when contradictory results occur. The result shows that both KPSS and ADF suffer from low power and size distortion and that the problems persists for the most common time series lengths. Problems particularly arise when the time series contains a trend or is a process with both an autoregressive and a moving average part. There is no clear evidence that one of the tests are superior to the other, it rather depends on sample size, parameter value and type of ARIMA process