On efficient optimisation of the CVaR and related LP computable risk measures for portfolio selection

The portfolio optimisation problem is modelled as a mean-risk bicriteria optimisation problem where the expected return is maximised and some (scalar) risk measure is minimised. In the original Markowitz model the risk is measured by the variance while several polyhedral risk measures have been introduced leading to Linear Programming (LP) computable portfolio optimisation models in the case of discrete random variables represented by their realisations under specified scenarios. Recently, the second order quantile risk measures have been introduced and become popular in finance and banking. The simplest such measure, now commonly called the Conditional Value at Risk (CVaR) or Tail VaR, represents the mean shortfall at a specified confidence level. The corresponding portfolio optimisation models can be solved with general purpose LP solvers. However, in the case of more advanced simulation models employed for scenario generation one may get several thousands of scenarios. This may lead to the LP model with a huge number of variables and constraints, thus decreasing the computational efficiency of the model. We show that the computational efficiency can be then dramatically improved with an alternative model taking advantages of the LP duality. Moreover, similar reformulation can be applied to more complex quantile risk measures like Gini’s mean difference as well as to the mean absolute deviation.

[1]  Helmut Mausser,et al.  Credit risk optimization with Conditional Value-at-Risk criterion , 2001, Math. Program..

[2]  D. Stoller,et al.  On the Generation of Normal Random Vectors , 1962 .

[3]  Philippe Artzner,et al.  Coherent Measures of Risk , 1999 .

[4]  R. Rockafellar,et al.  Optimization of conditional value-at risk , 2000 .

[5]  Georg Ch. Pflug,et al.  Scenario tree generation for multiperiod financial optimization by optimal discretization , 2001, Math. Program..

[6]  A. Müller,et al.  Comparison Methods for Stochastic Models and Risks , 2002 .

[7]  Wlodzimierz Ogryczak,et al.  On Extending the LP Computable Risk Measures to Account Downside Risk , 2005, Comput. Optim. Appl..

[8]  G. Pflug.  Some Remarks on the Value-at-Risk and the Conditional Value-at-Risk , 2000 .

[9]  Wlodzimierz Ogryczak,et al.  Dual Stochastic Dominance and Related Mean-Risk Models , 2002, SIAM J. Optim..

[10]  Wlodzimierz Ogryczak,et al.  Multiple criteria linear programming model for portfolio selection , 2000, Ann. Oper. Res..

[11]  J. Desrosiers,et al.  A Primer in Column Generation , 2005 .

[12]  J. Corcoran.  Modelling Extremal Events for Insurance and Finance , 2002 .

[13]  Maria Grazia Speranza,et al.  Conditional value at risk and related linear programming models for portfolio optimization , 2007, Ann. Oper. Res..

[14]  Hanif D. Sherali,et al.  Portfolio optimization by minimizing conditional value-at-risk via nondifferentiable optimization , 2010, Comput. Optim. Appl..

[15]  W. Ogryczak,et al.  LP solvable models for portfolio optimization: a classification and computational comparison , 2003 .

[16]  Wlodzimierz Ogryczak,et al.  From stochastic dominance to mean-risk models: Semideviations as risk measures , 1999, Eur. J. Oper. Res..

[17]  I. Maros.  Computational Techniques of the Simplex Method , 2002 .

[18]  C. Klüppelberg,et al.  Modelling Extremal Events , 1997 .

[19]  S. Yitzhaki.  Stochastic Dominance, Mean Variance, and Gini's Mean Difference , 1982 .

[20]  M. Rothschild,et al.  Increasing risk: I. A definition , 1970 .

[21]  Donald R. Barr,et al.  A comparison of multivariate normal generators , 1972, Commun. ACM.

[22]  H. Konno,et al.  Mean-absolute deviation portfolio optimization model and its applications to Tokyo stock market , 1991 .

[23]  S. Uryasev.  Probabilistic constrained optimization : methodology and applications , 2000 .

[24]  Maria Grazia Speranza,et al.  On LP Solvable Models for Portfolio Selection , 2003, Informatica.