Active Portfolio Model And Empirical Research Based On CVaR Total Risk Constraint

The mean-variance(MV) model which was put forward by Markowitz in 1952 laid the the core status of the finance portfolio theory.With the rapid development of financial market, more and more money in real investment activities separate its ownership and management, such as investment funds, investors usually entrust their money to fund managers, for the dual purpose of realizing the wealth’safety and good return depend on their professional knowledge, and at the same time investors will periodically evaluate managers’ performance. Then, the tracking error active profolio research is arised at the historic moment. As an improvement of mean-variance model, the tracking error model through the form of relative return and relative risk to scientific evaluate the performance of investment managers. In such a principal-agent relationship, fund managers will adopt active investment strategy to improve their performance based on the dual drives of benefit and pressure, while some risk-taking behaviors generated simultaneously will make the investors suffer a greater risk of loss.In addition, the unpredictable change of internal-external environment and financial market make the traditional model’disadvantages based on the assumption of certainty parameters and asset return distribution highlight increasingly, which aggravate the reasonable allocation of financial assets problem under uncertain environment become an urgent problem.Therefore, this article will further base on the perspective of risk control and uncertain environment to research the active portfolio management in the following aspects:Firstly, considering the limitations of variance or value-at-risk(Va R) risk measures, we introduce the conditional value at risk(CVa R) measuring the whole portfolio risk to improve Roll’s mean-tracking error variance model, which only focus on the tracking error while ignore the portfolio total risk. Meanwhile, for the consideration of investor’s investment tendency and market factors, multiple weight constraints and transaction cost alse introduced to establish an extented tracking error active portfolio optimization model which more in line with the reality of our country’ securities market. Finally the Shenzhen component index is used as an empirical object to analyse the practicability of extented model.Secondly, the fluctuation in financial market and changeableness in internal-external environment make the investment environment encounter many uncertainties, which casue the parameter estimation process relyed on historical data susceptible to estimation error interference. Therefore, this paper replace the traditional MV estimation method as an another Black-litterman model which combined market equilibriu m concept and investor’perceptions of future asset return to estimate model parameters, then establish an extented active portfolio optimization model under the BL model frame. Ten industry market indexes are also used for the comparative analysis of model’ performance under two parameter estimation methods.Finally, in the view of the above research, the robust optimization method is use to continuously analyze active strategies under uncertain investment cases. Based on Casta’s tracking error robust optimization model, this paper further introduce the uncertain asset information set and assumpt the asset return distribution belonging to a particular distribution set, then the CVa R method can be expand to the uncertain Worst-case CVa R. At the same time multiple weight constraints and practical transaction cost condition are introduced to build a robust tracking error model under WCVa R total risk constraint. Linear matrix inequality method is used to model’optimization and the Shanghai and Shenzhen 300 index is used as the empirical research object to analyse the practicability of extended model under certain an uncertain asset return distribution circumstances.This paper study the suitable and practical tracking error active portfolio strategies within the framework of traditional probability theory and robust optimization theory, and put forward the extension model under the restrictions of CVa R total risk and multiple weight constraints which solved successfully through computer software. This research not only can provide academic reference for the development of portfolio theory but also play a practical guiding significance for the investment decision of investment managers and the development of financial market.