by Neil Shephard, Sangjoon Kim
Reading guide
This entry refers to the 1994 Nuffield working paper by Sangjoon Kim and Neil Shephard. Its abstract describes a Gibbs-sampling framework for likelihood-based analysis of stochastic volatility, including maximum-likelihood and Bayesian estimation, a Monte Carlo comparison of estimators and comparison with ARCH models. A later 1998 journal version also credits Siddhartha Chib; keep the versions distinct when citing the work.
Review conditional variance and time-series estimation before comparing volatility models.
Build the conditioning, likelihood and inference foundations used in the paper.