Beyond index funds: data driven portfolio optimization for retail investors

This thesis examines whether data-driven portfolio strategies can improve long-term returnsfor retail investors compared to passive index investing. Using SEC data and Fama-Frenchfactor regressions as a basis, the study simulates small-cap value portfolios under real-worldconstraints such as taxes, transaction costs, and inflation. Monte Carlo simulations show thatmodest small-cap value tilts (e.g., 20%) can outperform broad market benchmarks if portfoliosare well-diversified and held long-term. However, style drift, implementation frictions, and thevariability of factor loadings limit consistency. The findings suggest that combining a low-costindex fund with a cautiously managed factor tilt offers a pragmatic strategy for retail investorsseeking higher returns without excessive risk.Keywords: factor investing, Small-Cap-Value, momentum, portfolio optimization, Monte Carlosimulation, retail investors, passive investingJEL Classification: G11, G17, C63, D14