A Stochastic Approach to Portfolio Optimization Using Competing Risk Metrics
First Faculty Advisor
portfolio construction models; empirical analysis; stochastic models; portfolio optimization; finance
This thesis presentation presents a stochastic approach to portfolio construction using various risk metrics as underlying models for portfolio optimization. The risk models utilized in this thesis include Mean-Variance, Minimum-Variance, Value-at-Risk (VaR), Conditional Value-at-Risk (CVaR). To evaluate the efficiency and overall performance of these models, historical data for 30 specific stocks was selected. The stock selection process focused on the selecting stocks that are highly volatile and correlated with one another. Empirical results reveal that portfolio optimization strategies outperform the benchmark. Additionally, results showed that the Minimum-Variance model constructed the best portfolio for the predetermined backtesting time period.