This paper establishes a necessary and sufficient condition for recurrence and ergodicity for the general first-order smooth threshold autoregressive processes. Depending on the tail behaviour of the smoothing distribution function, we give necessary and sufficient conditions under which the process is ergodic, null recurrent or transient.
|Number of pages
|Journal of Applied Probability and Statistics
|Published - 2012
- Nonlinear time series
- Null recurrent
- Smooth threshold autoregressive model