Institutional Portfolio Optimization

Efficient frontier and five allocation criteria on 25 EuroStoxx 50 stocks, with the final choice driven by VaR.

Asset Management · Course: Asset Management · 2025/26

Description

Simulates a portfolio manager’s decision process for a fund of European institutional investors. On 25 EuroStoxx 50 stocks and 10 years of monthly data: estimation of risk, return and correlation structure, construction of the efficient frontier and comparison of five allocation criteria under realistic concentration and sector constraints. The final portfolio is selected with a VaR criterion that favors downside protection and diversification over maximum Sharpe. Robustness tests and a backtest against the benchmark complete the work.

Skills

Portfolio construction, constrained optimization, estimation and reduction of estimation error, analysis of non-normal returns, backtesting.

Tools

Python (pandas, NumPy, SciPy/SLSQP, seaborn, hierarchical clustering), Jupyter, Excel.

Models

Markowitz (efficient frontier, tangency, GMV), Single-Index Model, Bayesian beta adjustment (Vasicek), Elton–Gruber cutoff, Kataoka safety-first, historical VaR, Jarque–Bera test, HHI index.

Educational projects. Any results are hypothetical, based on historical or simulated data, and are not indicative of future returns. Nothing on these pages constitutes investment advice or a solicitation to invest.

Next project: Pricing & Hedging Toolkit: Fixed Income and Derivatives

Open