A note on strong mixing of ARMA processes (Q1078909): Difference between revisions

Let \(\{Y_ t\}_ 0^{\infty}\) be an AR(1) process given by \(Y_ t=\rho Y_{t-1}+e_ t\) where \(| \rho | <1\) and \(e_ t\) are i.i.d. random variables independent of \(Y_ 0\). Let \(E[\{\log | e_ t| \}^+]<\infty\) and let \(e_ t\) have a nontrivial absolutely continuous component. The authors prove that then for any distribution of \(Y_ 0\), the process \(\{Y_ t\}_ 0^{\infty}\) is strong mixing. The result is generalized to AR(n) and to ARMA(n,m) processes.