Optimal Option Portfolios for Skew-Elliptical t Returns

Battle Count: 261

Highly relevant for quantitative portfolio managers dealing with options portfolios. The paper provides practical closed-form solutions for optimal option weights under realistic return distributions (skew-elliptical t). The findings on how skewness affects portfolio weights are directly applicable to risk management in options trading. However, the single-period framework and lack of trading costs limit direct implementation in high-frequency or dynamic trading strategies.

The analytical solutions (variance and CFVaR2) are tractable given the matrix formulas provided. However, implementing the full pipeline requires: (1) GARCH(1,1) filtering, (2) MLE fitting of multivariate skew-elliptical t-distribution, (3) computing Gosset Greeks via finite differences, (4) constructing the Q matrix and u vector with all intermediate variables, (5) numerical integration for option pricing, and (6) numerical optimization for CFVaR3. The tensor operations (Einstein summation) for the third moment add complexity.

The paper provides explicit analytical formulas for optimal weights (Theorem 3.1), uses publicly available data (yfinance), and references open-source tools (Scipy, NumPy, R SN package). However, no GitHub repository is provided, and some numerical optimization details (e.g., specific solver settings for CFVaR3) are not fully specified. The parameter estimation procedure follows Hu and Kercheval (2010) methodology.