Diversificação e eficiência: aplicação da análise envoltória de dados em rede na definição de portifólio de investimento no mercado acionário brasileiro

Abstract

Based on Data Envelopment Analysis and the portfolio management theory proposed by Harry Markowitz, this paper studies the efficiency in using these concepts to compose an optimal portfolio of stocks among the shares of companies listed on the Brazilian Stock Exchange (B3) in scenario of economic recovery post-crisis caused by Covid. It is founded that from economic data of the companies, allied to the macroeconomic and political conjuncture, the use of statistical resources and application of the Network Data Envelopment Analysis (NDEA), it is possible to construct a sample of papers whose diversified combination makes this investment portfolio optimal for the investor's profile. To carry out this assessment, a multiple case study was conducted, with an intentional sample of 140 companies registered at B3, distributed across 10 segments. The consolidated data for the third quarter of the year 2022 were considered, disclosed by each company in the sample. The variables used in this evaluation were two inputs (operating expenses and total assets), two intermediate inputs (net income and total revenues) and two outputs (average return on shares in the market and dividends). Then, the output-oriented NDEA was used to identify which companies are efficient to compose a stock investment portfolio. To measure performance and analyze the portfolio, the Ibovespa index, the main performance indicator for shares traded on B3, was chosen as a benchmark. The test resulted in a positive response to the hypothesis formulated regarding the possibility of success of this diversification method of choosing stocks even in an uncertain economic scenario. Finally, it concludes with the recommendation of the method, being an adequate strategy to assist investors in decision making. and that can be complemented by means of predictive resources, such as the bootstrap model.