Sieve Bootstrap for Strongly Dependent Stationary Processes

This paper studies the properties of the sieve bootstrap for a class of linear processes which exhibit strong dependence. The sieve bootstrap scheme is based on residual resampling from autoregressive approximations the order of which increases slowly with the sample size. The first-order asymptotic validity of the sieve bootstrap is established in the case of the sample mean and sample autocovariances. The finite-sample properties of the method are also investigated by means of Monte Carlo experiments.