A note on the asymptotic independence of the sum and maximum of strongly mixing stationary random variables (Q1897168)

Let \(\sigma^2_n= ES^2_n\to \infty\) \((n\to \infty)\) where \(S_n= \sum^n_{i= 1} X_i\) for a strictly stationary strongly mixing sequence \(\{X_i\}\) of centered r.v.s with finite variance. Let \(B_n\subset \mathbb{R}\) be Borel sets such that \(\liminf_{n\to\infty} P(C_n)> 0\) for \(C_n= \bigcap^n_{i= 1} (X_i\in B_n)\). If \(P(S_n/\sigma_n\leq x)\) converges to the \(N(0, 1)\) distribution function, then so does \(P(S_n/\sigma_n\leq x\mid C_n)\). This result implies asymptotic independence of \(S_n\) and \(\max_{1\leq i\leq n} X_i\), thus providing a generalization of a theorem by \textit{C. W. Anderson} and \textit{K. F. Turkman} [J. Appl. Probab. 28, No. 1, 33-44 (1991; Zbl 0726.60039)].