Convergence to type I distribution of the extremes of sequences defined by random difference equation

TL;DR

This paper investigates the extreme values of a sequence defined by a random difference equation, showing that under certain conditions, the normalized extremes converge to a double exponential distribution, extending previous results.

Contribution

It establishes convergence of the normalized extremes of the sequence to a double exponential distribution under mild conditions on the random coefficient M.

Findings

Normalized extremes converge to double exponential distribution

Convergence holds under mild conditions on M

Extends previous results to cases where P(M>1)=0

Abstract

We study the extremes of a sequence of random variables $(R_n)$ defined by the recurrence $R_n=M_nR_{n-1}+q$, $n\ge1$, where $R_0$ is arbitrary, $(M_n)$ are iid copies of a non--degenerate random variable $M$, $0\le M\le1$, and $q>0$ is a constant. We show that under mild and natural conditions on $M$ the suitably normalized extremes of $(R_n)$ converge in distribution to a double exponential random variable. This partially complements a result of de Haan, Resnick, Rootz\'en, and de Vries who considered extremes of the sequence $(R_n)$ under the assumption that $\P(M>1)>0$.