The effectiveness of fiscal policy in Brazil through the MIDAS Lens

Abstract

This paper aims to examine the effects of fiscal policy on economic growth in Brazil between 1999 and 2017. For this purpose, a novel methodology is applied, using a Vector Autoregressive with Mixed Frequency (MIDAS-VAR) model, proposed by Ghysels (2016), which allows for the estimation of the spending multiplier by measuring the impact of high-frequency data at low-frequency and vice versa. The impact of various types of spending on the gross domestic product (GDP) is analyzed, including primary expenditure, personnel, social benefits, subsidies, investment, and costing, while the central government’s primary revenue is used as the revenue variable. The expenditure and tax revenue are the high-frequency variables (monthly observations), whereas GDP is a low-frequency series (quarterly). The estimated fiscal multipliers for primary spending are less than one, suggesting no significant Keynesian effect on output, with particular attention given to the investment multiplier, which is estimated to be close to zero. Our results suggest that the frequency of data matters, and government expenditure has no significant impact on real GDP growth in Brazil. Therefore, the ability of Brazilian fiscal policy to influence economic growth may be limited.