Parametric portfolio policy with momentum-based sentiment trading strategy

Area of Science:

• Quantitative Finance
• Investment Management
• Behavioral Finance

Background:

• Conventional mean-variance portfolio models often lack adaptability to market dynamics.
• Integrating market sentiment and momentum can potentially improve asset selection and portfolio performance.
• Existing back-testing methods may not adequately account for real-world trading costs and dynamic market conditions.

Purpose of the Study:

• To develop and evaluate a parametric portfolio policy incorporating a momentum-based sentiment characteristic vector.
• To enhance the conventional mean-variance portfolio model by integrating market sentiment.
• To assess the model's performance against benchmark portfolios over an extended investment horizon, considering transaction costs.

Main Methods:

• Development of a parametric portfolio policy using a momentum-based sentiment characteristic vector.
• Optimization of interrelationships between portfolio holdings, benchmark portfolio, and characteristic vectors.
• Evaluation of the model using a 15-year investment horizon with historical rebalancing data from S&P 500 and Dow Jones 30 stocks, accounting for transaction costs.

Main Results:

• The proposed model significantly outperforms benchmark models, including the minimum-variance model.
• The sentiment-augmented momentum strategy demonstrates superior performance compared to the conventional mean-variance approach.
• During financial crises, the model's increased selection of sentiment-based momentum leads to distinct asset allocations and potentially higher investor utility.

Conclusions:

• Integrating market sentiment into characteristic vector construction is crucial for enhancing portfolio performance.
• Parametric portfolio policies offer significant value in improving asset allocation and risk-adjusted returns.
• The sentiment-augmented momentum strategy provides a robust alternative to traditional portfolio optimization methods.