Análise de determinantes no sucesso da reorganização financeira de empresas em recuperação judicial

Abstract

The field of bankruptcy and insolvency is an area that is still in development, with many studies proposing new bankruptcy prediction models to bring greater accuracy to financing decisionmaking for banks and investment for shareholders. This research addresses the topic of bankruptcy and judicial recovery, and it is structured in the different articles. The first is a systematic literature review article, in which we sought to identify the evolution of the research field and its state of the art, as well as identify, through qualitative analysis, the types of models and explanatory variables most used in bankruptcy prediction. The main results in this article were that the most commonly used model for bankruptcy prediction in the analyzed studies was Altman's Z-Score; the most used variables are profitability, liquidity, and debt ratio; and the most common types of models were logistic regression and multivariate discriminant analysis. The research continues with the second article, which is a quantitative study applying Z-Score models to Brazilian samples, aiming to analyze both the models' performance and the variables. For this article, it was found that the Altman models did not perform as well in the selected sample as they did in the sample for which they were developed, achieving a 10% Type I error and a 33% Type II error for the Revised Z-Score model, and a 52% Type I error and a 3% Type II error for the Emerging Markets model. However, based on the results of the F-tests applied to the Z-Score models' variables, they have significant discriminant abilities for the bankruptcy and judicial recovery event. The third and final article is the development of an emergency or judicial recovery prediction model, which, unlike most existing studies in the field, does not aim to identify when a company will become insolvent or bankrupt, but rather to identify among those in this state which are capable of recovering. In this work, the proposed model achieved an AUC of 95.74, a 17% Type I error, and a 4% Type II error. For practical purposes, this research provides relevant information for decision-making by managers and investors, not only for bankruptcy but also for corporate recovery and emergence, and identifies significant variables for this event.