New Approaches To Portfolio Optimization Using Drawdown To Measure Risk Aversion

The current economic scenario stems from many uncertainties generated by the pandemic period and global political and social issues, making it essential for investors to compose better investment portfolios. This paper describes an optimization model that minimizes the risk of an investment portfolio, where the risk is the conditional value-at-risk of the series of drawdowns, subject to a minimum acceptable return constraint by the investor. This model is equivalent to a well-known model in the literature, which will be referred to as the original model. The main objective of this paper is to enhance the original model by presenting new versions, which will henceforth be called new models. The new models, like the original model, were computationally implemented in JULIA language with Ipopt. The results of the numerical experiments were analyzed and demonstrated that, computationally, they are more efficient, thus enabling them as alternatives to the original model, especially for problems involving large volumes of data. The new models discussed produced greater financial efficiency compared to the Brazilian Ibovespa index, and additionally, a way of using the model as a tool for risk control of financial investments in the Brazilian market is presented.