PYXIDA Institutional Repository
and Digital Library
Collections :

Title :Optimal asset allocation using CVaR
Creator :Daktylidi, Iliana
Contributor :Topaloglou, Nikolaos (Επιβλέπων καθηγητής)
Athens University of Economics and Business, Department of Accounting and Finance (Degree granting institution)
Type :Text
Extent :95p.
Language :en
Abstract :In our study, we present and analyze some categories of risk that affect theportfolio management. We cite and analyze the most known risk metrics used by assetmanagers. We discuss the important and complementary role of derivatives to theportfolio management and we also analyze the variety of derivative products. Wedevelop an integrated simulation and optimization framework for optimal assetallocation using the optimization model Conditional Value-at-Risk (CVaR). Our aimis to assess the performance of this optimization model as risk management tool inselecting diversified portfolios composed of stocks on the FTSE/ATHEX 20. In ourempirical research, we are referred to one market which is the Greek market, settingas benchmark the index of FTSE/ATHEX 20. The simulation is applied to domesticportfolios that consist of stocks on the FTSE/ATHEX 20. Referring to the descriptivestatistics of our data, we recognize that they are not normally distributed and presentassymetric return distributions. Therefore, we develop and implement the model thatoptimize the CVaR. At first, we use the CVaR model for a static test, deriving theefficient CVaR-return frontier for each potential investor, for the single period of01/2014. Thereafter, through backtesting over the period of 01/2011-12/2013 andusing historical data, we investigate empirically the ex post performance of theunhedged CVaR portfolios on stocks of FTSE/ATHEX 20. We also compare theperformance of the CVaR unhedged portfolios against the CVaR hedged portfoliosthat incorporate futures contracts. Hedging using futures proves to be the superiorstrategy that improves the performance of the unhedged CVaR portfolios. The latterincur the market risk exposure and thus, the unhedged portfolios follow a downwardtrend with losses corresponding to that of FTSE/ATHEX 20. Although in static teststhe CVaR model selects portfolios on the efficient frontier curve for every kind ofinvestor (risk averse, risk neutral and risk lover), it does not succeed in an efficientdiversification according to the backtesting experiments. This happens becausedomestic securities present high correlations between their returns and the systematicmarket risk remains, reducing the portfolio value. Over the backtest period and usingthe CVaR model, we form three different ex post portfolios (aggressive, defensiveand middle) one for each kind of investor. We conclude that the inclusion of futurescontracts into the three different portfolios contributes to an outstanding andprofitable performance for all the three portfolios, hedging the market risk andincreasing the portfolio value above the initial budget of each investor.
Subject :Portfolio management
Conditional Value-at-Risk
Date :30-09-2017
Licence :

File: daktylidi_2017.pdf

Type: application/pdf