Stochastic Volatility Models Predictive Relevance for Equity Markets

Abstract

This paper builds and implements multifactor stochastic volatility models where the main objective is step ahead volatility prediction and to de-scribe its relevance for the equity markets. The paper outlines stylised factsfrom the volatility literature showing density tails, persistence, mean reversion, asym-metry and long memory, all contributing to systematic data dependencies. As a by-product of the multifactor stochastic volatility model estimation, a long-sim-ulated realization of the state vectors is available. The realization establishes a functional form of the conditional distribution, which is evaluated on observed data convenient for step ahead predictions. The paper uses European equity for relevance arguments andillustrational prediction purposes. Multifactor SV mod-els empower volatility visibility and predictability enriching the amount of infor-mation available for equity market participants.