Search

The objective of this study is to perform an econometric modeling exercise of the individual series of taxes aiming to obtain income elasticity and the future projection for each tax. For this, we apply dynamic linear models (MLD) and dynamic factor (MFD), both estimated based on the Bayesian approach. Our sample of taxes accounts for over 90% of the Brazilian tax burden, consisting of monthly data between December 2006 and May 2019. The forecast within the sample is made for the period of one year while the projection out of the sample is from June 2019 to June 2020. The results obtained corroborate the expectation regarding the adequacy of the methodologies used. The exercises for forecast validation showed excellent performance considering the various evaluation criteria. In particular, for most cases, the percentage accumulated error was below 4%; and in some cases the value obtained for this indicator was below 1%.

The goal of this article is to model the desagregated series of taxes in Brazil. We use monthly data of a sample of taxes in charge for 80% of the Brazilian gross tax burden in the 1995-2010 years. To estimate the model we employ a Dynamic Linear Model (DLM) with variable parameter (West & Harrison 1997). The choice of this particular model was motivated by the constant changes made in the Brazilian tax system during these years. The conditional forecast is performed a year ahead out of the sample. The main conclusions of the paper are the following. In general the results seem strongly satisfactory. The forecasts fall inside the error bands and the predicted error is bellow 10% until six steps ahead. Above this horizon the forecast lose efficiency. Although the model performed quite well for some taxes, further efforts are required for others. Finally, the elasticity appears to have fluctuated below the unity or the majority of taxes O objetivo deste estudo é modelar as séries individuais para uma amostra de tributos que corresponde por cerca de 80% da carga tributária bruta brasileira. Para isso, usou-se um modelo linear dinâmico bayesiano com parâmetros variáveis. A aplicação deste modelo tem como justificativa as alterações sucessivas ocorridas no sistema tributário nacional que podem implicarem mudanças nas elasticidades relevantes. Os resultados obtidos corroboraram a expectativa quanto à adequação desta metodologia. De um modo geral, temos que os valores observados para a previsão condicional fora da amostra ficaram dentro do intervalo de confiança, sendo que o erro de previsão não se situou acima de 10% nos seis primeiros meses. Os valores das elasticidades na maioria dos casos ficou abaixo da unidade.

Este artigo apresenta uma metodologia de estimação da carga tributária bruta brasileira (e de alguns de seus principais componentes) em níveis trimestrais, detalha os problemas práticos enfrentados na implementação dessa última (por volta de abril de 2008) e, por fim, apresenta as estimativas obtidas para o período que vai do primeiro trimestre de 1995 até o quarto trimestre de 2007.This article presents a methodology for estimating quarterly figures for the Brazilian aggregate gross tax burden (and for some of its main components), details the practical problems one has to tackle when trying to implement it (as of April, 2008) and, finally, presents the estimates one obtains for the period beginning in the first quarter of 1995 and finishing in the fourth quarter of 2007.