MODELO PREDITIVO PARA GERAÇÃO DE INFORMAÇÕES GERENCIAIS PARA EMPRESAS PATROCINADORAS DE ENTIDADES FECHADAS DE PREVIDÊNCIA COMPLEMENTAR.

The objective of this study is to carry out a comparative analysis with the application of two models, the Autoregressive Integrated Moving Average (ARIMA) and the Vectorial Autoregression (VAR), in the sponsoring companies of the Entidades Fechadas de Previdência Complementar, listed in Brasil, Bolsa, Balcão (B3). Through a quasi-experimental action, two models were utilized, the Autoregressive Integrated Moving Average (ARIMA) and the Vector Autoregression model (ARV). Based on the predictive results of these models, it was possible to identify that the data run by the ARIMA method did not present a good fit, it was possible to conclude that the model created by ARV was more robust in predicting future situations of the analyzed companies, and that, decisions based on predictive models, pointed to the need to anticipate decisions on the discount rate, the inflation rate, the rate of salary increase and on the guaranteeing assets. [ABSTRACT FROM AUTHOR]