Estratégia de alocação para portfólios de crédito corporativo no Brasil utilizando Data Envelopment Analysis (DEA)

• Ano de defesa: 2021
• Autor(a) principal: Cunha Júnior, Décio
• Orientador(a): Eid Júnior, William
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Debêntures; Otimização de portfólio; Modelo de Markowitz; Carteira com máxima diversificação
• Palavras-chave em Inglês: Data Envelopment Analysis (DEA); Corporate bonds; Portfolio pptimization; Markowitz’s model; Portfolio with maximum diversification
• Link de acesso: https://hdl.handle.net/10438/30649

The purpose of this thesis is to use the Data Envelopment Analysis (DEA) methodology for corporate credit portfolios in Brazil, specifically made up of corporate bonds (debêntures); however it can be extrapolated to products with similar characteristics. Optimization of credit portfolio of corporate bonds is treated in very few academic papers and the use of DEA for this purpose is unprecedented in the country. This research consists in optimizing portfolios by checking the efficiency of the methodologies of Markowitz and Data Envelopment Analysis. Three analyses will be carried out, the first using the Markowitz model for a period resulting in an efficient frontier portfolio that presents maximum diversification; the second, using two classic DEA models; and the last, through the application of the efficient frontier model to a portfolio with maximum diversification in the two DEA models, resulting in a hybrid methodology. The reference prices for the debentures are provided by a self-regulator in the financial market. Criteria to inputs and outputs are defined regarding the DEA model. This modeling considers as inputs and outputs variables, respectively, the volatility of the spread and the duration of the asset, the return on the credit spread and the current rating of the bond. Twenty-four monthly allocations are performed and, at the end of each period, the results were compared in each allocation of the respective models. As a result, the portfolios generated by DEA methodology were performed better than the set of assets generated by Markowitz’s model for a portfolio with maximum diversification.