Combining bootstrap methods: a Monte Carlo experiment

The idea of this work is to study whether the combination of different bootstrap methods can lead to an improvement in the performance, as it does in the forecasting framework where is widely used. Given that it represents a rather challenging set-up, we will focus on long memory time series and we will explore the combination of three very well-known bootstrap methods for time series with long range dependence. We present a very preliminary Monte Carlo experiment that provides interesting and promising results.