Revisit the SETAR model for the 3-month Treasury Bills in Section 16.2.1 using the updated time series from Exercise 2. Does the SETAR model proposed in Section 16.2.1 hold? Change the value of the threshold from 8.5% to 8%, 7.5% and 7%. In these new models, is it still possible to find two regimes, stationary and nonstationary? Are any of these models better than the model with the 8.5% threshold? Comment on your findings.