On the relationship between beta-Bartlett and Uhlig extended processes

TL;DR

This paper explores the relationship between beta-Bartlett and Uhlig extended stochastic volatility processes, showing their similarities and differences in Bayesian inference, and introduces a backward sampling algorithm for beta-Bartlett.

Contribution

It demonstrates the close relationship but non-equivalence of the two processes and provides a new backward sampling method for beta-Bartlett.

Findings

Models can have identical posteriors and forecasts but differ in joint distributions.

Bayes factors may not distinguish between these models, requiring alternative comparison methods.

Introduces a backward sampling algorithm for beta-Bartlett process.

Abstract

Stochastic volatility processes are used in multivariate time-series analysis to track time-varying patterns in covariance matrices. Uhlig extended and beta-Bartlett processes are especially convenient for analyzing high-dimensional time-series because they are conjugate with Wishart likelihoods. In this article, we show that Uhlig extended and beta-Bartlett are closely related, but not equivalent: their hyperparameters can be matched so that they have the same forward-filtered posteriors and one-step ahead forecasts, but different joint (smoothed) posterior distributions. Under this circumstance, Bayes factors can't discriminate the models and alternative approaches to model comparison are needed. We illustrate these issues in a retrospective analysis of volatilities of returns of foreign exchange rates. Additionally, we provide a backward sampling algorithm for the beta-Bartlett process, for which retrospective analysis had not been developed.