摘要：Monte Carlo simulations are used to study the size and power properties of two stationarity tests developed by Kwiatkowski et al. (1992) [KPSS] and Leybourne and McCabe (1994) [LMC] when the data contain additive outliers. We show that the KPSS tests are very robust to additive outliers whereas the LMC test exhibits size distorsions and loss of power.
