A note on the asymptotic normality of sample autocorrelations for a linear stationary sequence (Q1817515): Difference between revisions

The author considers a stationary process \(\{X_t\}\) given by \(X_t= \sum^\infty_{k=-\infty} \psi_k Z_{t-k}\), where \(\{Z_t\}\) is a strictly stationary martingale difference white noise. Let \(f(\lambda)\) be the spectral density of \(\{X_t\}\). Under the conditions that \(\int f^2(\lambda) d\lambda<\infty\) and \(m^\tau \sum_{|k|\geq m}\psi^2_k\to 0\) for some \(\tau>1/2\), it is proved that the sample autocorrelations are asymptotically normal. This is an extension of a theorem previously published in the literature in which \(\tau=1\) was assumed. The present result has an application in long memory models.