Ordinal- and Continuous-Response Stochastic Volatility Models for Price Changes: An Empirical Comparison

Abstract

AbstractOrdinal stochastic volatility (OSV) models were recently developed and fitted by Müller & Czado (2009) to account for the discreteness of financial price changes, while allowing for stochastic volatility (SV). The model allows for exogenous factors both on the mean and volatility level. A Bayesian approach using Markov Chain Monte Carlo (MCMC) is followed to facilitate estimation in these parameter driven models. In this paper the applicability of the OSV model to financial stocks with different levels of trading activity is investigated and the influence of time between trades, volume, day time and the number of quotes between trades is determined. In a second focus we compare the performance of OSV models and SV models. The analysis shows that the OSV models which account for the discreteness of the price changes perform quite well when applied to such data sets.KeywordsMarkov Chain Monte CarloPrice ChangeCredible IntervalStochastic VolatilityTrading ActivityThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.