Investigation of the volatility spillover effect and dynamic conditional correlations in Tehran Stock Exchange using wavelet based Bayesian conditional variance heteroscedasticity

Characteristic of skewness, fat tails and frequency dimension are important features of financial time series that have not been taken into account in classical econometric models. Therefore, in this study, the Bayesian method for conditional variance heteroscedasticity based on wavelet analysis has been used to investigate the volatility spillover effect and dynamic conditional correlations in three sub-periods between the daily return data of selected Tehran Stock Exchange (TSE) indices during the period from December 14, 2008 to April 20, 2019. Sub-periods are defined according to the Joint Comprehensive Plan of Action (JCPOA), which includes the pre-JCPOA period, JCPOA period, and the period after United States withdrawal from the JCPOA. The results of the Bayesian DCC GARCH (1,1) model, with the rejection of the constant conditional correlation hypothesis versus the dynamic conditional correlation hypothesis based on posterior marginal distribution in all subsections, indicates that the impact of shocks on the volatility of stock returns in wavelets and sub-periods are not the same. Also, Bayesian dynamic conditional correlation graphs are recommended for each sub-periods and in each wavelet, a different stock for a suitable investment. The main idea used in this research can be used in the calculations of asset risks and the selection of optimal asset portfolios.