The DEA method and the traditional approaches in the selection of Brazilian stock funds
Efficiency; Investment Funds; Markowitz; DEA; Bootstrap.
This study aims to evaluate portfolios assembled with stock funds with active index, using DEA methods, obtaining better performance than those obtained with traditional methods. For this, an analysis was carried out using Data Envelopment Analysis (DEA) method, with Variable Returns to Scale (VRS) and Constant Returns to Scale (CRS) and product orientation, in its classic form, as well as in together with the Bootstrap tool for more robust results. Subsequently, Markowitz optimization was carried out. There was the adoption of the analysis in three distinct stages. First, the efficiency coefficients for the classic DEA models and for the DEA model with bootstrap were verified. In sequence, the formation of the efficient portfolio projected from the efficient funds by the DEA models was carried out. Then, Markowitz optimization was performed for the projected efficient portfolios and a classical portfolio. The database comprised the period between 2012 and 2021, with data extracted from the Economática®. Among the main results, it was identified that the optimization of the classical portfolio obtained better results than those estimated by the DEA models. It is also verified that exogenous factors such as the recession period and the pandemic caused by Covid influenced the results.