1

I was wondering if there is a "rule of thumb" range (or something more concretely) for AR coefficients that would indicate whether or not a process is strongly mean reverting, mean reverting, random walk and etc. Any advice would be greatly appreciated. For example: let's say my AR coefficients for the given levels in a AR(1) quantile regression (assume the lines are not crossing) are:

  1. 0.05 = 1.02
  2. 0.15 = 0.95
  3. 0.50 = 0.72
  4. 0.85 = 0.25
  5. 0.95 = 1.5

Needless to say the process acts as a random walk at 0.05 and 0.95 but what about the remaining ranges?

  • Certainly does not look like a coding question. Suggesting migration to stats.stackexchange.com – IRTFM Oct 15 '15 at 00:44
  • How exactly do I transfer this question to stat.stackexchange.com? (first time doing it) –  Oct 15 '15 at 17:47
  • I think you can request it yourself by "flag"-ing and asking that the moderator do it. – IRTFM Oct 15 '15 at 17:49

0 Answers0