Abstract: In this paper, based on Bayes factor and credible interval methods, we deal with the unit root test for time series with heavy distribution. Monte Carlo simulations demonstrate the validity of the Bayes factor approach and the credible interval approach in this paper. We compare and analyze these two methods. In addition, we consider the effect of prior information and degree of freedom on unit test results. At last, these two methods are applied to the time series of the unemployment rate and consumer price index in the United States, and we find that there are unit roots in these time series.

Key words: unit root test, prior density function, Bayes factor, credible interval