Portfolio optimization using variance-adjusted mean and predictive structures
01 Dec 2020
Reading time ~1 minute
This project records my role as co-advisor for the undergraduate thesis “Otimização de um Portfólio de Investimentos Utilizando o Modelo de Média Ajustada pela Variância e Estruturas Preditivas” by Lucas Lobo Ataide and Pedro Vieira Kluppel Carrara at the University of Brasilia.
The thesis starts from a practical criticism of using arithmetic return in the Markowitz portfolio model and proposes a variance adjustment to better estimate the portfolio’s true return. From there, it adds a predictive structure based on financial indicators, a leverage policy inspired by one quarter of the Kelly criterion, and a risk analysis using Monte Carlo, Value at Risk, and correlation.
The abstract concludes that the proposed model improves adjusted return estimation and shows gain potential above the benchmark used in the study.
In the PDF acknowledgements, Lucas Ataide explicitly thanks his co-advisor Breno Rodrigues Brito for the effort invested in helping build the work. That mention appears on PDF page 5.