The objective of this study is to verify the macroeconomic effects of government spending shocks on activity in Brazil from 1999 to 2017. For this purpose, we calculate the spending multiplier using a methodology not yet applied in the literature, a Vector Autoregressive with Mixed Frequency, proposed by Ghysels (2016), more specifically, a mixed data sampling VAR - MIDAS-VAR. The mixed-frequency VAR models allow us to measure the impact of high-frequency data at low frequency and vice versa. To calculate the impact of the increase in spending on GDP (gross domestic product), different types of spending were considered, namely: primary expenditure, personnel, social benefits, subsidies, investment, and costing. The central government’s primary revenue was used for the revenue variable. The expenditure and income series are the high-frequency variables (monthly observations), whereas GDP is a low-frequency series (quarterly). The fiscal multipliers of primary spending found here are less than one, which suggests that there is no so-called Keynesian effect on output. Among the spending components, the investment multiplier attracted the most attention because its estimate was close to zero.
Comissão Organizadora
Anderson Odias da Silva
Claudia Yoshinaga
Ricardo D. Brito
Felipe Saraiva Iachan
Vinicius Augusto Brunassi Silva
Complete tool for submission, evaluation and publication of proceedings for scientific events.
