Multiperiod Mean-CVaR Portfolio Selection
Due to the time inconsistency issue of multiperiod mean-CVaR model, two important policies of the model with finite states, the pre-committed policy and the time consistent policy, are derived and discussed. The pre-committed policy, which is global optimal for the model, is solved through linear programming. A detailed analysis shows that the pre-committed policy doesn’t satisfy time consistency in efficiency either, i.e., the truncated pre-committed policy is not efficient for the remaining short term mean-CVaR problem. The time consistent policy, which is the subgame Nash equilibrium policy of the multiperson game reformulation of the model, takes a piecewise linear form of the current wealth level and the coefficients can be derived by a series of integer programming problems and two linear programming problems. The difference between two polices indicates the degree of time inconsistency.
Keywordsmean-CVaR pre-committed policy time consistency in efficiency time consistent policy linear programming integer programming
Unable to display preview. Download preview PDF.
- 4.Björk, T., Murgoci, A.: A general theory of Markovian time inconsistent stochasitc control problem, working paper (2010)Google Scholar
- 8.Gao, J.J., Zhou, K., Li, D., Cao, X.R.: Dynamic Mean-LPM and Mean-CVaR Portfolio Optimization in Continuous-time, working paper, arXiv:1402.3464 (2014)Google Scholar
- 9.Rockafellar, T.R., Uryasev, S.P.: Optimization of Conditional Value-at-Risk. Journal of Risk 2, 21–41 (2000)Google Scholar