Curva de phillips novo-keynesiana no Brasil: uma nova abordagem econométrica

Abstract

This research aims to advance the standard New Keynesian Phillips curve (CPNK) estimation methodology proposed by Gali & Gertler (1999) in Brazil. Therefore, a new aproach was used, suggested by Barnichon & Mesters (2020), which consists of identifying the NKPC through monetary shocks, in addition to constructing the confidence interval of the parameters by inverting the AR test with Almon restriction, which is robust to weak instruments. Two estimates of the standard NKPC were performed: unrestricted and restricted. The second consists of the sum of the parameters of the components backward-looking and forward-looking is equal to one, which implies a vertical standard NKPC in the long run. For the inertial and expectation parameters, the results obtained are in line with the available literature, and without major differences in the two estimates made, with both positive impacts on the current inflation rate. However, with the methodological difference, evidence was found that the previous ones, in general, overestimate the inflationary inertia and underestimate the forward-looking component. With regard to the output gap, they pointed to a null impact on the current inflation rate, corroborating part of the NKPC studies for Brazil. Thus, despite the standard NKPC for long-term verticality, since the restriction imposed in the second estimation is on the twodimensional confidence interval of the parameters backward-looking and forward-looking, it is rejected this hypothesis.