The Research Of Portfolio Optimization Based On CDaR

The research of portfolio optimization is the base of risk prevention and stability of financial market. The key of portfolio optimization is to measure portfolio risk, and the development of portfolio optimization is meanly about measuring of risk. Markowitz proposed Mean-Variance model, and researchers proposed VaR, CVaR and models based on these later. CDaR and its portfolio optimization model are proposed by A.chekhlov in2000in order to strengthen the effect of portfolio optimization. We chose daily price of10stocks from May11,2009to March3,2010as the study sample. The mean work and conclusions can be stated as follows.(1) Obtain efficient frontier of Mean-CDaR model by Matlab7.10. The shape of efficient frontier is a curve tilt from lower left to top right. This feature is consistent with the law of efficient frontier of portfolio.(2) Portfolio constitution of Mean-CDaR model. In the portfolio, the proportion of stocks which have lower means of daily rate of return and higher CDaR is high, otherwise, the proportion is low. This result demonstrates the effective of Mean-CDaR model.(3) Nature of efficient frontier of Mean-CDaR model. The efficient frontier will move right with the increase of confidence level, with the decrease of weight constraint, the efficient frontier will move right and the maximum expected rate of return will become less.(4) Comparison of Mean-VaR model, Mean-CVaR model and Mean-CDaR model. We compare the three models use direct method and map method. The conclusion is that Mean-CDaR model is more comprehensive in measure portfolio risk and have better portfolio optimization effect.