Mudanças de Estado e Multiplicadores Fiscais no Brasil entre 1999-2012

• Autor(a) principal: Samuel, Marco Antonio Castelo Branco
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UERJ
• Texto Completo: http://www.bdtd.uerj.br/handle/1/7616
• topic: Fiscal Multipliers; Fiscal Policy; MS-SBVAR Methodology; Markov Switching; Bayesian Structural VAR; Multiplicadores Fiscais; Política Fiscal; Metodologia MS-SBVAR; Markov Switching; VAR Estrutural Bayesiano; Política tributária - Brasil; Política monetária; CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA

This dissertation evaluates the behavior of fiscal multipliers in Brazil from 1999 to 2012. It uses a methodology developed by Sims, Waggoner e Zha (2008), which is a Bayesian estimation procedure that allows for state (regime) dependent endogenous change in model s parameters and the states follow a markovian process of regime change. It estimates a Structural Bayesian VAR model with Markov Switching regimes (MS-SBVAR). The database comprises the consumption of public administration, the fixed capital gross formation of the public administration, the net tax burden and the Gross Domestic Product (GDP) of the three levels of government (federal, state, including the Federal District, and municipalities). The software MATLAB / Dynare was used to estimate the model and the results suggest the occurrence of 2 or 3 regimes in the two best data fitting models. The different estimated multipliers show the correct signs and, as expected, they are higher for exogenous shocks to public administration s fixed capital gross formation which are effective, since they have values higher than one and long-term impact on GDP - when compared with exogenous shocks to public administration s consumption, which have a small persistence and are ineffective (less than one), and a negative and persistence response of GDP to shocks in net tax burden. The results do not also show a higher fiscal multiplier in regimes with higher model s residuals variance.