portfolio optimization using multivariate garch models: evidence from tehran stock exchange

in this paper, in order to optimize the portfolio consisting of selected industrial stocks of petroleum products, automobiles and parts, electrical industry and extraction of minerals from tehran stock exchange member, first, time – varying conditional covariance matrix has been estimated based on the following multivariate garch models: diagonal-vech (1,1), ccc (1,1) and diagonal -bekk (1,1). then the portfolio risk minimization problem has been solved according to the markovitz theory and the optimal time-varying weights of the four selected industries have been specified. optimization results indicate that based on all above models, more weight in the portfolio allocated to the industries that their volatility in stock returns is less than others. the optimal time-varying weights for industries that have increase in their output volatility, has fallen and vice versa if the output volatility over time has decreased, and the share has increased.