Eficácia da política fiscal no Brasil: uma abordagem SVAR identificado com restrições de sinais e de zeros

Abstract

The objective of this study is to assess the effectiveness of fiscal policy in Brazil, including whether its impacts are significantly affected when it does not alter the government's fiscal deficit or when different periods of time are considered. For this purpose, it was adopted a Structural Bayesian Vector Autoregression model with an innovative identification, proposed by Arias et al. (2014) in another context, which not only imposes signal constraints, but also zeros constraints on the Impulse Response Functions. This new way of identifying fiscal innovations allows avoiding ambiguities in the identification of shocks. As there is no consensus in the theoretical and empirical literature on the effects of fiscal policy, it is important to measure them under alternative hypotheses of identification. The results indicate a change in the GDP response in Brazil in the two different time periods considered: in the first period of time (1999-2007), a reduction of taxes and increase of expenses increases the GDP; in the second period analyzed (2009-2016), the GDP response is the opposite. In relation to Fiscal Multipliers that do not affect the government budget, a positive GDP response was obtained for government consumption shocks, while for government investment shocks there was a change in the GDP response between the different time intervals, due to the form of the other variables studied reacted to shock.