Stochastic volatility model’s predictive relevance for Equity Markets

EasyChair Preprint no. 2339

Abstract

This paper builds and implements multifactor stochastic volatility models. The main objective is volatility prediction and its relevance for equity markets. The paper outlines stylised facts from volatility literature showing density tails, persistence, mean reversion, asymmetry and long memory, all contributing to systematic dependencies. Applying long simulations from stochastic volatility (SV) models and filter volatility using a form of nonlinear Kalman filtering, the unobservables of the nonlinear latent variables can be forecasted with associated fit characteristics. The paper uses European equity data from United Kingdom (Ftse100) and Norway (Equinor) for relevance arguments and illustrational prediction purposes. Multifactor SV models seem to enrich volatility predictions empowering equity market relevance.

Keyphrases: Bayesian estimators, Kalman filter, M-H algorithm, MCMC Simulations, stochastic volatility

@Booklet{EasyChair:2339,
  author = {Per Bjarte Solibakke},
  title = {Stochastic volatility model’s predictive relevance for Equity Markets},
  howpublished = {EasyChair Preprint no. 2339},

  year = {EasyChair, 2020}}