Modeling the Daily Volatility of Oil, Gold, Dollar, Bitcoin and Iranian Stock Markets: An Empirical Application of a Nonlinear Space State Model

Using the daily data of returns of 9 assets during the period from 04/29/2013 to 09/20/2021, the volatilities of different asset markets were modeled in this paper. The multivariate factor stochastic volatility model (MFSV) under the nonlinear space-state approach provides the basis for decomposing asset return volatility into two components, “volatility rooted in latent factors” and “idiosyncratic volatility”, and for estimating the time-varying pairwise correlation of time series. The results show: First, there are three latent factors so that the volatility of returns of five Iranian stock markets is affected by the first and third hidden factors, while the volatility of the other four international markets is mainly affected by the second latent factor. Second, the idiosyncratic volatility of the different Iranian stock returns exhibits clustering behavior, and there is a relatively strong correlation among them. Third, the volatility of oil returns is explained by the hidden factors, and consequently their idiosyncratic volatility is almost smooth. Fourth, the correlations between the return volatility of bitcoin and the volatilities of other conventional assets are negligible. The results of this paper may be useful for future research on investment opportunities and risk-return characteristics of portfolios.