Robust conditional value–at–risk under parallelpipe uncertainty: an application to portfolio optimization

Kara, Güray
In markets with high uncertainties, the trade–off between maximizing expected return and minimizing the risk is one of the main challenges in modeling and decision making. Since investors mostly shape their invested amounts towards certain assets and their risk version level according to their returns; scientists and practitioners has done studies on this subject since the beginning of the stock markets’ establishment. Developments and inventions in the mathematical optimization provide a wide range of solutions to handle this problem. Mean–Variance Approach by Markowitz is one the oldest and best known approaches to the risk–return trade–off in the markets. However, it is a one time–step model and not very much prepared for highly volatile markets. After Markowitz, different optimization approaches have been invented for portfolio Optimization, especially, in the tradition of Conditional Value–at–Risk. In this study, we modeled a Robust Optimization problem based on the data and used Robust Optimization approach to find a robust optimal solution to our portfolio optimization problem. This approach includes the use of Robust Conditional Value–at–Risk (RCVaR) under Parallelpipe Uncertainty sets, an evaluation and a numerical finding of the robust optimal portfolio allocation. We obtained and then traced back our robust linear programming model to the Standard Form of a Linear Programming model; then we solved it by a well–chosen algorithm and software package. The main idea is modeling a robust portfolio optimization problem that includes our development of RCVaR based on uncertainty–set–valued data. Our aim is, by considering the return–risk trade–off analysis under uncertain data, to obtain more robust, in fact, lower, risk–level under worst–case scenario by using RCVaR. Uncertainty in parameters, based on uncertainty in the prices, and a risk–return analysis are crucial parts of this study. Hence, the trade–off (antagonism) between accuracy and risk (variance), and robustness are our main issue. A numerical experiment is presented containing real–world data from stock markets. The thesis ends with a conclusion and an outlook to future studies. 


Stability advances in robust portfolio optimization under parallelepiped uncertainty
Kara, Guray; Ozmen, Ayse; Weber, Gerhard Wilhelm (2019-03-01)
In financial markets with high uncertainties, the trade-off between maximizing expected return and minimizing the risk is one of the main challenges in modeling and decision making. Since investors mostly shape their invested amounts towards certain assets and their risk aversion level according to their returns, scientists and practitioners have done studies on that subject since the beginning of the stock markets' establishment. In this study, we model a Robust Optimization problem based on data. We found...
Empirical comparison of portfolio risk diversification algorithms
Yerli, Çiğdem; Kestel, Sevtap Ayşe; Schindler, Nilüfer; Department of Financial Mathematics (2018)
The enhanced correlations during global financial crisis has revealed that simple asset allocation portfolios prove to be not well-diversified across different risk factors, which makes the risk based asset allocation strategies popular. However, the strategies still construct the risk concentrated portfolios due to the correlation among the asset classes. As a result, risk allocation among uncorrelated risk factors instead of risk allocation among asset classes have become widely used. This thesis aims to di...
Value-based risk management in defense projects pricing with a robust optimization approach
Aker, Kerime Özge; Meral, Fatma Sedef; Department of Industrial Engineering (2015)
In this study, we address the bid pricing problem of a defense company facing with supply chain and financial risks due to long-term project planning. Creating shareholder value has become a quite popular topic in business as value-adding companies are the only ones to survive in the market. We formulate the problem as a value-based performance and risk management model through robust optimization approach in order to cope with various ambiguities over project lifecycle with possible discrete scenario set. ...
A Comparison of constant and stochastic volatility in Merton’s portfolio optimization problem
Öztürk, Ozan; Sezer, Ali Devin; Department of Financial Mathematics (2018)
Merton's Portfolio Problem is a dynamic portfolio choice problem, which assumes asset returns and covariances are constant. There is well documented evidence that, stock returns and volatilities are correlated. Therefore, stochastic volatility models in dynamic portfolio problems can give better results. The work [J. Liu, Portfolio selection in stochastic environments, Review of Financial Studies, 20(1), 2007] developed a general dynamic portfolio model that allows the parameters of the model to depend on a...
Uncertainty assessment for the evaluation of net present value: a mining industry perspective
Erdem, Özge; Demirel, Nuray (2012-05-01)
The investment decisionmaking process in the insurance and finance industries is affected by new advances such as simulation techniques. These advances have improved the discounted cash flow method (DCF). In DCF, a dynamic and flexible model construction is not possible because it does not consider uncertain conditions. Each project should be evaluated taking into account the related uncertainties because the related uncertainties determine the characterization of the project. Related uncertainties can be p...
Citation Formats
G. Kara, “Robust conditional value–at–risk under parallelpipe uncertainty: an application to portfolio optimization,” M.S. - Master of Science, Middle East Technical University, 2016.