Stable limit distributions for strongly mixing sequences (Q909333)

Let \(\{X_ j:\) \(j\geq 1\}\) be a stationary strong mixing sequence of random variables. The paper establishes a necessary and sufficient condition for the convergence of the normalized partial sums of \(\{X_ j\}\) to a nonnormal strictly stable limit distribution with index p; when \(p\in [1,2)\) it is assumed that the partial sums are symmetric. No mixing rate is required. The proof uses the theory of the independent case through a blocking argument. A theorem giving sufficient conditions for this behaviour (with symmetry and \(\rho\)-mixing with a certain mixing rate assumed when \(p\in [1,2))\) is derived which is related to theorems of \textit{R. A. Davis} [Ann. Probab. 11, 262-269 (1983; Zbl 0511.60021)] and of \textit{A. Jakubowski} and \textit{M. Kobus} [J. Multivariate Anal. 29, 219-251 (1989)]. This improves a result of the reviewer [Ann. Probab. 12, 390-426 (1984; Zbl 0542.60012)] for the \(\phi\)-mixing case. Concerning general limits in that case see also \textit{H. A. Krieger}, Isr. J. Math. 47, 32-64 (1984; Zbl 0536.60032) (which improves \textit{H. Bergström}, Periodica Math. Hungar. 2, 173-190 (1972; Zbl 0252.60009)). In the \(\psi\)-mixing case, rates of convergence to stable limits were obtained by \textit{L. Heinrich} [Math. Nachr. 131, 149-165 (1987; Zbl 0631.60020)].