O efeito de variáveis macroeconômicas nacionais nas receitas públicas: um estudo para a previsão do ICMS dos estados do nordeste

Abstract

This dissertation aims to present an alternative methodology for forecasting tax revenues in the states of the Northeast, managed by their Secretariats of Finance. For this, a linear regression model was used to project the most important state own tax revenue, the Tax on Circulation of Goods and Services (ICMS), based on the partial short-term effects of national macroeconomic variables, such as the GDP, inflation, basic interest rate and exchange rate. The study will be divided into four main parts: review of state revenues and taxes in Brazil, revenue forecasting methodologies, research methodology and presentation of results and discussions. The results obtained were satisfactory, so that it was possible to forecast the annual ICMS revenues of the Northeastern states with average absolute forecast errors between 1.48% and 6.83%. The results are important so that it is possible to estimate, within an acceptable degree of predictability, the ICMS revenues of the Northeastern states for the following year through the Annual Budget Law.