An Empirical Study Of Portfolio Optimization Based On CVAR Method

Risks arising from the uncertainty of future outcomes, as well as investors, the limited nature of their knowledge, Financial Investment Risk factors specific to such uncertainty caused by changes in equity prices, the possibility of losses to investors, as well as the extent of losses.As the risk of potentially harmful and therefore must be effective monitoring, the risk comes to minimize the loss of it may lead to and from the risk to get the appropriate returns, effectively manage risk. Risk management is the most important steps in risk analysis and assessment, that is, measurement of risk, select what kind of indicators to measure risk is essential.Major research portfolio theory is that in the case of uncertain future outcomes, how the assets of the limited kinds of choices and combinations, making the expected return and risk achieve a reasonable balance.In 1952, Markowitz first introduced a scientific portfolio selection theory:a rational investor is always looking at a given level of returns expected under the conditions of minimum risk, or at a given level of risk the largest portfolio returns since then, to the expected return to measure the rate of portfolio returns in order to measure the variance or standard deviation of portfolio risk mean-variance model in the portfolio area of research has been established.However, with the securities market and the size of the expansion of securities trading and financial theory and financial engineering development, securities investment risk measurement methods have become more integrated and complex. The traditional measure of risk-based approach to variance by a number of academics have also questioned as to VaR as the representative of modern risk measurement approach has become the world’s most popular tool for risk control standards. In order to overcome the lack of VaR, conditional VaR CVaR risk measurement techniques have been academics considered to be a VaR risk measure than the more rational and effective techniques of modern risk management practices.In this paper, mean-variance based on the theory of optimization problems, will minimize the variance of the objective function to promote the CVaR as a measure for the risk of a new portfolio optimization model. In this paper, the goal is to establish a complete model of the computer system. Through the model, can be considered by the selection of financial assets out of CVaR values calculated or predicted the highest accuracy models to be used for measurement of financial assets.In China, the current CVaR study focused on analysis methods. Using of various distribution assumptions and estimates of the volatility of financial time series models, to solve the peak-thick tails and volatility clustering phenomenon. Compared with the analysis methods,Monte Carlo is a full-value estimation method assumes that market factors do not obey normal distribution, analytical method can effectively solve the problems in the non-normal difficulties encountered. However, the financial time series of peak-thick tails and volatility clustering phenomenon, the general Monte Carlo in the calculation of CVaR when there are limitations.In order to enhance the performance of the estimated CVaR with Monte Carlo, improve estimation accuracy, Comprehensively understand of the Monte Carlo.In the research,I used t distribution and the GARCH (1,1) volatility estimation model applied to the Monte Carlo, by and use our country’s financial market data, empirical tests to the fund as a study to compare theoretical model portfolio and fund the size of the actual investment portfolio risk and to measure the use of Sharp, by comparison between the performance of priority bad. So that a mean-CVaR model used in the Chinese market as well as exploring the feasibility of China’s financial markets, the mean-CVaR model choice based on calculation method.