Multi-move sampler for estimating non-Gaussian time series models: Comments on Shephard and Pitt (1997)

This note points out a problem in the multi-move sampler proposed by Shephard and Pitt (1997) and corrects their method. The performance of the original Shephard and Pitt (1997) method and the correct method is examined by estimating stochastic volatility models using simulated data. It is found that the original method yields an estimation bias which increases with the number of blocks while the correct one performs well irrespective of the number of blocks.