Previsão do ICMS do Ceará a partir dos CNAES e índice de incerteza de arrecadação

Abstract

Most of the resources in the state budget of Ceará have tax origins. The most important tax for the state treasury is the Tax on the Circulation of Goods and Transport and Communication Services (ICMS). Forecasting this revenue is vitally important for executing a balanced public budget. Therefore, aiming to develop a methodology to improve the prediction of state revenue, this study verifies whether models of Autoregressive Vectors with Dynamic Factors, which include several macroeconomic variables from Brazil and Ceará, together with data from ICMS collection by CNAE, improve the forecasts of total ICMS collection in the State of Ceará compared to ARMA-type time series models. The results show that the FAVARs models using CNAES improve forecasts of ICMS collection in Ceará when compared to the ARMA univariate models. The study concludes that the FAVAR model, using CNAE tax collection data, presents solutions in term of mean squared errors for predicting the tax collection of the best ICMS in the state of Ceará.