Perfil dos aplicadores de poupança do Estado do Ceará através de análise multivariada de dados

Abstract

The present work aims to create a model that best represents the estimated balance of financial reserves of people considering some of their individual characteristics, as well as to discriminate which variables best represent this model. The variables used were age, gender, income, city of residence and the existence or not of other investments. The study is restricted to the analysis of the profile of savers in the state of Ceará, for which uncharacterized registration information was collected from Caixa Econômica Federal savings account holders who reside in the state. The choice of the asset to be analyzed was based on recent studies presented by the Brazilian Association of Financial and Capital Market Entities - ANBIMA, which shows that the preferred investment of Brazilians is the savings account. The analysis was performed using a multivariate data analysis technique called discriminant analysis. We took advantage of this and performed the modeling in two periods, July 2019 and July 2020, in order to verify whether there were changes in the savings patterns of the Ceará population as a result of the health crisis caused by COVID-19. The results presented show that the basic information collected by banks in the process of opening a savings account produced a discriminant function capable of correctly classifying 36.1% of the samples in the year 2020 and 36.7% in the year 2019, besides that, it is not identified, up to the moment of this study, significant changes in the characteristics of the saver brought about by the effect of the pandemic.