Investors have to search for an equilibrium between portfolio risk and return during their decision-making process. One of the first and most common mathematical models for this problem was formulated by H. Markowitz [1], where the variance of the portfolio
return is used as a risk measure. However, this measure has several serious disadvantages. An alternative measure, Conditional Value-at-risk (CVaR), has become very popular and widely used because it has all properties a risk measure should have. Unfortunately, both optimization frameworks suffer from being sensitive to small variations in the input parameters. One of the widely used and effective approaches for optimization under parameter uncertainty is a robust optimization.
The major objective of this thesis is to compare different frameworks, such as Markowitz and mean-CVaR optimization models, as well as its robust counterparts, for multi-asset class portfolios. The results show that only in the case of non-normal return distributions
there is a difference between Markowitz and mean-CVaR optimal portfolios. The simulation study also shows that robust counterparts have significantly smaller risk in both
models, losing at the same time not so much in portfolio return. It means that portfolio robustification can be a good choice for the conservative investor.
«
Investors have to search for an equilibrium between portfolio risk and return during their decision-making process. One of the first and most common mathematical models for this problem was formulated by H. Markowitz [1], where the variance of the portfolio
return is used as a risk measure. However, this measure has several serious disadvantages. An alternative measure, Conditional Value-at-risk (CVaR), has become very popular and widely used because it has all properties a risk measure should...
»