On the Performance of Mean-Risk Portfolio Optimization Strategies

Finding a portfolio strategy that entails optimal performance and risk diversification may be a complicated task for investors. In this thesis, we explore and evaluate the performances of several non-trivial portfolio optimization strategies, based on various measures of risk, to identify the optimal choice. Furthermore, as in contemporary papers, we will also assess whether or not trivial strategies are outperformed by those that rely on rigorous theoretical frameworks. These portfolio strategies were backtested on historic U.S. stock marketdata ranging from 2000 to 2015, to evaluate how they had performed in reality at the time. In accordance with similar studies, the performance of a trivial equally weighted strategy was not significantly distinguishable from the non-trivial counterparts analyzed in this paper. However, it required 42% more reallocations,possibly implying higher operational costs. By contrast, a non-trivial Expected Shortfall strategy performed better in general, particularly considering its Sharpe ratio. Based on this and mainly its appealing theoretical properties, it is deemed the best strategy to pursue in portfolio management