Abstract
The least square estimator (LSE) of three-regime TAR models has been well developed when the noise has a finite variance. However, little is known about the asymptotic behavior of the LSE of three-regime threshold autoregressive (TAR) models in the heavy-tailed setting and neither theory nor methodology is available for model fitting in this case. In this paper, we use simulation method to study the asymptotic behavior of the LSE of infinite variance three-regime TAR models. A real example is given for the NASDAQ daily volume from the year 2012 to 2014. The results in this paper may provide an empirical insight for further theoretical research of the LSE of three-regime TAR models with heavy-tailed innovations.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
