Persistence of AR(1) sequences with Rademacher innovations and linear mod 1 transforms

arXiv2305.10038MaRDI QIDQ6510149FDOQ6510149

Authors: Vladislav V. Vysotsky, Vitali Wachtel

Abstract: We study the probability that an AR(1) Markov chain

X_{n + 1} = a X_{n} + x i_{n + 1}

, where

a i n (0, 1)

is a constant, stays non-negative for a long time. We find the exact asymptotics of this probability and the weak limit of

X_{n}

conditioned to stay non-negative, assuming that the i.i.d. innovations

x i_{n}

take only two values

p m 1

and

a l e f r a c 23

. This limiting distribution is quasi-stationary. It has no atoms and is singular with respect to the Lebesgue measure when

f r a c 12 < a l e f r a c 23

, except for the case

a = f r a c 23

and

m a t h b f P (x i_{n} = 1) = f r a c 12

, where this distribution is uniform on the interval

[0, 3]

. This is similar to the properties of Bernoulli convolutions. It turns out that for the

p m 1

innovations there is a close connection between

X_{n}

killed at exiting

[0, i n f t y)

and the classical dynamical system defined by the piecewise linear mapping

x m a p s t o f r a c 1 a x + f r a c 12 p m o d 1

. Namely, the trajectory of this system started at

X_{n}

deterministically recovers the values of the killed chain in reversed time! We use this property to construct a suitable Banach space, where the transition operator of the killed chain has the compactness properties that let us apply a conventional argument of Perron--Frobenius type. The difficulty in finding such space stems from discreteness of the innovations.

This page was built for publication: Persistence of AR($1$) sequences with Rademacher innovations and linear mod $1$ transforms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6510149)