Aplicação da Técnica de Análise das Demonstrações Contábeis em empresas brasileiras e o esboço de um Termômetro de Insolvência

Abstract

The development of efficient predictive methods to prevent losses in investments or in company health analysis becomes increasingly critical, due to the plurality of types emerging business in the current decade. For this reason, this research was designed to to improve the models of management, controllership and risk analysis in the concession of credit. For this, a qualitative and quantitative research was developed using data financial statements of Brazilian companies from the last 4 years, made available by the Securities - CVM, together with the São Paulo Stock Exchange - BOVESPA, applying statistical analysis techniques, such as Discriminant Analysis, specifically, the Multiple Linear Regression, in order to elaborate an outline of a statistical function that could predict the situation of the companies, in this case, allocating them in situations of solvent or insolvent. After surveying the financial indicators using a hierarchical theoretical method, regression was tested to verify that the prerequisites of the statistical method were met. The result of the research demonstrated that the generated model did not fulfill all the prerequisites necessary to make it viable, but it presented progress and potential for a study from this model, since this result presented, in a general analysis of the procedures, a degree of significance.