Using Bayesian Inference for Stochastic Volatility Model(Empirical Evidence from Damascus Securities Exchange)

Abstract

Volatility is especially important in financial markets, as the value of stocks and financial instruments depends on their risks, which are explained by volatility. Therefore, this paper proposes a fully Bayesian Simulation, based approach for statistical inference in stochastic volatility models, the workflow in the research is included analysis and modeling of the daily returns of the Damascus Securities Exchange during the period from (1/4/2010) to (12/9/2021), and using of three types of models based on different assumptions about volatility (ARIMA - GARCH - BSV). The direct in-sample comparison between the three models indicated that the Bayesian Stochastic Model performs better in terms of predictive accuracy and in clarifying the characteristics of the data used in the analysis.  The results of evaluating out-of-sample predictions using the Bayesian Stochastic Model also showed a convergence between the actual values ​​and the estimated values ​​of returns, with their movement in the same direction. These results show the importance of using Bayesian Inference for both researchers and investors in the financial markets due to its superiority over the widely applied models (ARIMA - GARCH).