The Analysis And Application Of The Portfolio Model Based On CVaR

Judging from China's actual situation, during the transition in the economic system in China has accumulated a large number of non-performing assets, become China's economic operation of the destabilizing factors, at the same time, China has also emerged a number of financial institutions, bankruptcy or shut down, on China's economic operations produced a certain amount of adverse effects, domestic financial market has shown unprecedented volatility in。In addition, as in recent years,, further liberalization of China's financial markets, this volatility will continue to continue and there is growing trend.As the financial system itself the inherent vulnerability, determine the financial risk is inherent and inevitable. Ability to conduct effective risk management has been considered the financial industry, and even the regulatory authorities, including whether a country's financial system, competitiveness and sustainable development capacity of the key. Risk measurement is the basis of financial market risk management and core. Whether it is generated risk analysis report, to hedge or decentralized approach to the transfer of risk, or to determine risk-adjusted capital limits. First, the market had to risk the size and the possibility of the risk of measurement. The qualities of risk measurement, to a large extent determine the effectiveness of risk management in the financial markets. Can be said that no accurate measurement of risk, market risk management impossible. When the investment institutions, entrepreneurs or financiers onto the international stage and found that they face a greater risk, have been working hard to find risk-averse, risk diversification of the channels and means, so a number of risk measurement and management methods should be transported and birth. This risk through the identification, measurement, decision-making and monitoring, risk reduction brought about uncertainty and ultimately to optimize the allocation of resources in order to reduce the risk of purpose. This article is based on the above topics of the context.The following brief summary of the important risk under the portfolio model idea:in 1952 raised Markowitz mean-variance model (MV model). Markowitz on investment approach can be understood as four stages. The first stage, a group of investors to develop the securities as an investment object; The second stage, the prospect of those securities to investors, that all studies estimate the expected rate of return of securities, the standard deviation and covariance between the securities; The third phase, to determine the feasible set and efficient frontier; The fourth phase, to determine the optimal portfolio investors. Markowitz model of portfolio investment theory, laid the cornerstone of Hyundai Securities. In its theory, portfolio risk is measured by varianceδp2. MV model is very important to a hypothetical portfolio of the assets of the joint distribution of the yield of normal distribution. However, many scholars'research and practice indicate that the rate of return normality assumption is that distortion, MV model and therefore has been questioned. The 20th century, the nineties, there VaR methods, and quickly, widely accepted by numerous financial units. VaR is the meaning of an asset or a portfolio at a given confidence level and holding period, due to market fluctuations caused by the negative asset or combination of the maximum potential loss. Mathematical formula is:Prob(ΔV> VaR)= 1-β,whichβfor the confidence level.ΔV holding period for securities portfolio, in the loss of At. VaR of the biggest advantages is its quantitative standardization, thereby creating a unified framework to all financial institutions to quantify the risk of the portfolio as a simple figure, easy to understand high-level managers and regulatory agencies. The most commonly used method of calculating VaR historical simulation method, analytical method and Monte Carlo simulation method. However, when applied to the VaR portfolio, we will find that set up the mean-VaR model would create two key mistakes. First, VaR is not a consistent risk measure, while the study indicated that only consistent risk measurement methods in order to act as a portfolio management tool. Consistency includes four aspects:subadditivity; positive homogeneity; monotone; pass invariance. VaR does not satisfy subadditivity, that it is not a convex risk measure, that is, the portfolio risk measure will not necessarily be less than the portfolio and single asset risk measure. In addition, VaR is not a sufficient measure of loss tail. It leads to neglect the small probability of occurrence of large losses and even the financial crisis, which is precisely the risk management, must be concerned about.To overcome the shortcomings of VaR, Rockafeller and Uryasev, in 1999, this new proposed CVaR risk measurement method. CVaR is the Conditional Value-at-Risk of the abbreviation, usually translated into terms of value at risk may also be known as the average excess of loss (Mean Excess Loss),the average shortfall or tail VaR (Tail VaR), its meaning could be interpreted as a certain confidence level (confidence level), with losses exceeding VaR of the conditional mean, reflecting the excess of loss average. It answered, "what in the current market factors in the allocation of resources makes the investment portfolio, as well as the minimum value of the minimum expected loss (the VaR value) is exceeded when the amounts of loss?" Issue. It better reflects the portfolio VaR than the potential risk, CVaR risk measurement method has a wide range of applications, such as credit risk measurement, the determination of internal risk capital, capital allocation, and financial supervision. This article focuses on the study is based on CVaR portfolio model model analysis and application.In this paper, the conditional value at risk (CVaR) portfolio model analysis and application of the theme, a more comprehensive study of the CVaR system, including the basis for the introduction of CVaR advantages and disadvantages of VaR, both in the field of application of the portfolio and contrast, The final analysis of CVaR in China's securities market in the use of problems to be solved, that is, give us challenges. Research has focused on:1 Non-normal conditions, the mean-CVaR efficient frontier portfolio model.2 Constraints on the mean-CVaR efficient frontier portfolio model effects.3 Non-normal conditions, the mean-CVaR efficient frontier model and the MV model differences.This paper is divided into six chapters.Chapter I describes the research background, significance and methods.Chapter II conditions value at risk (CVaR) portfolio of domestic and international literature review, respectively, research from the CVaR approach, CVaR risk measures and other methods of comparative study, CVaR used in the portfolio, CVaR in other areas of the use of four aspects of summarized under the relevant domestic and international literature review.Chapter III comparative study of CVaR and VaR, first introduced to the definition and commonly used VaR several ways, through the relevant literature summarized the advantages and disadvantages of VaR, VaR is a standard statistical techniques used to measure the risk of a comprehensive and integrated approach than other subjective, artistic, strong traditional risk management approach to more accurate reflection of the risks faced by financial institutions the situation, has greatly increased the scientific nature of risk management systems, but the findings have shown a good experience, too simple VaR risk measurement approach has serious shortcomings, the major defects of it are as follows:(1) its incompliance with the thermo of consistency; (2) the non-sufficiency of the measurement of the tail loss. The average-VaR model is enlisted to notify the weaknesses of VaR. After analyzing the shortcomings of VaR is proposed to improve the conditions of value at risk (CVaR) of the definition and the corresponding calculation method to analyze the process of CVaR in the calculation of the choice of relevant parameters, such as:confidence level, holding period, as well as the application of CVaR, there are four aspects:(1) investment portfolio optimization, (2) to determine internal risk capital requirements and setting risk limits, (3) allocation of capital, (4) information disclosure and financial regulation. CVaR and VaR through comparative study, CVaR advantages are mainly two points:(1) CVaR is to satisfy subadditivity; (2) CVaR measure the tail of the loss is full.Chapter IV Mean-CVaR model theory analysis, Mean-CVaR mainly conducted a theoretical analysis, in this paper that a "good" portfolio optimization model we should have a perfect theory of nature, but also easy to solve, expand and implementation. Based on CVaR portfolio optimization model has the following advantages:(1) CVaR recognized effects of decentralization; (2) CVaR model gives the optimal solution is an effective second-order random accounts; (3) CVaR model is usually convex, and can effectively avoid the multi-extremum problems; (4) in limited scenarios, CVaR model can be designated as a linear problem, and thus easily solved with the expansion of large-scale investment is suitable for solving combinatorial optimization problems; (5) CVaR as a downside risk measure, suitable for for solving options and other derivatives containing the portfolio optimization problem. Through the above analysis, first of all to build mean-CVaR model, respectively discussed the continuous and discrete loss of function f(x, y) in both cases, the model solution, get in when the loss function f(x,y) is continuous, can be directly through the derivation of the confidence interval, solution CVaR model; when the loss function f(x,y) is a discrete, CVaR model by linear programming solutions. Then analyzes the case of a single period of no friction mean-CVaR portfolio optimization model and the corresponding expansion of models, including transaction costs, investment ratio constraints, liquidity constraints, and industry or plate constraints such as constraints.Chapter V mean-CVaR model of empirical research, by empirical research and analysis mean-CVaR model, in this chapter:(1) Comparative analysis of the mean-CVaR model and under the MV efficient frontier of different models, and comparative analysis of each map. (2) the mean-CVaR model, confidence and transaction costs on the mean-CVaR models. Through the above two studies, this chapter the following conclusions:(1) in the non-normal conditions, the mean-CVaR model and the MV efficient frontier model are very different, and when a gradual increase in confidence when the gap increases. (2) The expected portfolio returns M and the confidence level is the same, the mean-CVaR model is greater than the corresponding MV corresponding, MV model the corresponding CVaR is greater than the mean-CVaR model for the corresponding CVaR. (3) With the continuous increase confidence in the transaction cost is fixed, the mean-CVaR efficient frontier model and investment in the feasible set to the right movement; the same time, the expected portfolio returns as M increases, the same degree of confidence in the model CVaR value increases. (4) With the transaction costs increase, the confidence level is fixed, the efficient frontier shift to the right, the same combination with the expected rate of return CVaR corresponds to the value of M increases, and as the transaction costs increase, the portfolio efficient frontier of lowering the ceiling.Chapter VI Conclusions and recommendations, pointed out that the application of CVaR in China face problems and suggestions for improvement, while noting that this lack of presence and future research directions.