周期异方差时间序列的季节单位根检验

Tests for Seasonal Unit Roots in Periodical Heteroscedastic Time Series

为检验带异方差的季节时间序列中的单位根,提出基于最小二乘估计的统计量.在原假设下得到检验统计量的极限分布.用M on te C arlo方法计算同方差下的经验分位数并考虑异方差对检验水平的影响.实例分析表明了用该方法检验季节单位根的有效性.

We propose test statistics based on least square estimators for unit root with periodical heteroscedastic shocks in seasonal time series. Under null hypothesis, the limit distributions of statistics are obtained. The percentiles under hemoscedasticity are given and the effect of periodical heteroscedasticity on the test sizes is considered by Monte Carlo method. A real example shows the validity of this test method.