Parameter Estimation questions

Hi Everyone,

I would like to ask if it is okay to have the 95th percentile of the posterior distribution of AR(1) persistence to equal to 1.

Thank you,
Alex

1 Like

No, that is a sign of trouble, typically of having a data mean inconsistent with the model steady state and forcing the model to have a unit root to explain the different mean.

Thank you Prof. Pfeifer. I appreciate your help.