"Constraints And Equivalence In Cvar Portfolio Optimization" by Casey Rozowski

Date of Award

1-1-2012

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Economics & Finance

First Advisor

Daniel Biederman

Abstract

The implementation of constraints are standard practice for a portfolio manager. The effect, mathematical and practical, is a study still in development. This thesis offers a standardized approach to finding a potential explanation of these constraints. It considers this approach in both the MVO and CVaR settings both theoretically by providing a mathematical proof, and practically by simulating portfolio optimizations with both generated and actual market data. In the MVO setting an MVO with constraints is equivalent to an unconstrained MVO with perturbed covariance matrix. With regards to CVaR, the CVaR optimization with constraints is equivalent to an unconstrained CVaR optimization with a perturbed asset returns matrix. While this study clarifies how this process can be applied to other portfolio optimizations under constraint it leaves room for a deeper study of how the adjustment of particular individual constraints can effect portfolio optimization.

Share

COinS