Comparação entre modelo de regressão e séries temporais para a arrecadação do IPVA em Sergipe

Abstract

The Tax on Motor Vehicles (IPVA) was developed through the draft Law 804/85, to replace the Highway Single Rate (TRU), established in 1969, linked to costs of the transport system. In recent years, the collection of this tax has shown high growth, driven by the financial and fiscal measures that stimulated the increase in vehicle sales. This study seeks to analyze the behavior of this tribute and develop models that will fit this tax revenue for the state of Sergipe. In order to adjust this model, the study uses econometric methods, specifically temporal and regression series that fit the behavior of the collection of property taxes. The historical series was obtained from the Secretary of State of Sergipe Finance and covers the period from January 2000 to June 2015. To model the series, the met hodologies of Box e Jenkins, Linear Regression, Quadratic and Exponential. The choice of the final model was based on minor errors relative. The best model was the SARIMA (0,1,1)(1,1,0)12, which was more accurate than the others, therefore being the best fit for the series, which has better forecasts.