BTS is a re-sampling method used to simulate the empirical distribution of a statistic (Effron and Tibshirani, 1993). The same method is used to evaluate the null hypothesis of no change between the two periods (e.g., pre-payment and postpayment periods).