When Markowitz Meets Ben Graham – Portfolio Optimization vs Judgment

1-line summary

I compared a real equity manager’s portfolio to a Markowitz-optimized allocation to understand where quantitative efficiency ends and fundamental judgment begins.

Overview

I analyzed an actively managed equity portfolio and compared its weights to those generated by Markowitz mean–variance optimization. The goal was to understand why investors override models, and where optimization fails to capture business quality, valuation, and downside protection.

Project Motivation

In theory, Markowitz gives the best portfolio for a given risk. In practice, investors override it. I wanted to understand why, and where models provide discipline versus where they miss what matters in long-term investing.

Technical Details

I reconstructed the manager’s portfolio, computed returns, vol, and correlations, built an efficient frontier, and generated a max-Sharpe portfolio. Then I compared optimized weights stock by stock against the manager’s weights and evaluated the business reasons behind differences.

Results and Insights

The optimized portfolio delivered higher theoretical efficiency, but removed positions a long-term investor might intentionally hold. Markowitz optimizes based on the past, investors allocate based on how businesses evolve. The strongest outcome came from blending both approaches.

Conclusion

Quant models are powerful tools for risk control, but they are not decision-makers. Real investing sits between mathematical discipline and fundamental conviction.