Mean-Earnings-Variance based Portfolio Selection

The objective of the thesis is to provide a portfolio selection methodology which will take into account the earnings of the companies whose shares are under consideration, though a look into the problem of estimation of the efficient frontier for a basket of stocks. Following an overview of standard portfolio selection theory, a simple risk-return based model will be expanded to a return-earnings-risk model and the portfolio frontier will be projected as a surface in a third dimensional portfolio space. An inquiry into the optimal estimation method will seek to provide the methodology that will lead to the combination of stocks with the greatest Sharpe ratio. The expanded model is then back tested against the standard one employing data for Dow Jones components for the period between 1997 and 2008. Looking closer into these results, it will be shown how the inclusion of earnings lead to better portfolio risk weighted performance, as measured by the Sharpe ratio.