Incorporating statistical distribution in loss and gain functions in CVaR robust mean-variance portfolios

A new approach to optimize a portfolio of financial instruments by minimizing conditional value-at-risk (CVaR) incorporating statistical distributions into gain and loss functions is presented. Our proposed model has less number of variables and constraints as well as the ability of covering different ideas about the priorities of optimization, determined as a multiplier of gain in our objective function. Since current conditional value at risk model can be obtained from our model, the new model can be considered as the expansion of previous one. We demonstrate how the portfolios of our model can be considerably more profitable in comparison with current model. To do so, the datasets of three assets from Iranian stock exchange between 22nd December and 4th July are used to find optimum portfolios and the next 90 days for testing the real performances.