One Year of Inflation Targeting in Brazil Implementing
One Year of Inflation Targeting in Brazil Implementing Inflation Targeting in Brazil Joel Bogdanski Alexandre Tombini Sérgio Ribeiro da Costa Werlang
One Year of Inflation Targeting in Brazil Macroeconomic Models 4 Instrumental for managing monetary policy under IT 4 Powerful tool for communicating monetary policy (inflation fan charts)
One Year of Inflation Targeting in Brazil Challenges for Macro Modeling 4 Exchange-rate passthrough 4 Endogeneization of exchange-rate movements 4 Forward vs. Background-looking Phillips curve 4 Role of inflation expectations 4 Role of prices set by the public sector
One Year of Inflation Targeting in Brazil Building Blocks Demand (IS Curve) 4 Supply (Phillips Curve) 4 l l l 4 Exchange-rate (UIP) l 4 exchange-rate passthrough forward x backward looking inflation expectation endogenous x exogenous risk premium Interest rate rules l l l Taylor type rules predetermined path optimal rules
One Year of Inflation Targeting in Brazil 4 Non Fiscal IS 4 Fiscal IS Demand Side (IS curve) where: h log of output gap. r log of (one plus) real interest rate Pr log of (one plus) total primary deficit /GDP h, hf white noise.
One Year of Inflation Targeting in Brazil 4 Backward-looking 4 Forward-looking 4 Combined Supply Side (Phillips Curve) where: h p. F e Et(. ) b, f, n log of one plus inflation. log of output gap. log of foreign producer price index log of exchange rate. Expectation on time t. white noise.
One Year of Inflation Targeting in Brazil Modeling the passthrough where: p. F log of foreign producer price index. e log of exchange-rate. E exchange-rate.
Treatment of inflation expectations One Year of Inflation Targeting in Brazil 4 Forward-looking Phillips curve 4 Alternatives l Institutional approach
One Year of Inflation Targeting in Brazil Treatment of inflation expectations 4 Alternatives l Model Consistent (recursive solution)* where a b means that b is in a neighborhood of a. * - The convergence is usually achieved in less than 20 iterations.
One Year of Inflation Targeting in Brazil 4 Exchange-rate determination rate follows a UIP: where: e i. F x log of exchange rate log of foreign interest rate log of risk premium residual including the expectation variations assumed white noise
One Year of Inflation Targeting in Brazil 4 Exchange-rate determination Modeling the risk premium l l exogenous path endogenous determination 3 depends on PSBR/GDP ratio (primary) and other risk premium determinants. where: X risk premium (SOT) in basis points PR PSBR/GDP ratio (primary) Zj other risk premium determinants
One Year of Inflation Targeting in Brazil 4 Taylor Interest rate rules type rules where: log of inflation * log of inflation target h log of output gap i log of interest rate degree of interest rate smoothing ( = 1, conventional Taylor rule) ’s arbitrarily set or obtained through an optimization procedure 4 Predetermined l l path fixed nominal rate (fan chart) budget trajectory
One Year of Inflation Targeting in Brazil 4 Optimal l l Interest rate rules Non-stochastic simulation: find an interest rate path that minimizes the following loss-function. Stochastic simulation: find an interest rate path that minimizes the following loss function. This simulation is more computer demanding than the non-stochastic one.
One Year of Inflation Targeting in Brazil 4 Forecasting Scenarios l Model specification 3 l Exogenous variables 3 l Copom defines which relations are relevant for the monetary policy decision. The most likely path for the exogenous variables are set by the Copom after interacting with the staff. Shocks 3 The timing, magnitude, variance and skewness are set by the Copom after interacting with the staff.
One Year of Inflation Targeting in Brazil 4 Forecasting Fan Chart l Measure of central tendency 3 l Shocks stylization 3 l median: the model estimate the mean, median is obtained using the variance and skewness of a two-piece normal distribution. The magnitudes are obtained from out of model estimation. The assessment of variance and skewness are subjective. Variance 3 It is calculated using the historical forecast error as benchmark. However, it can be adjusted by subjective assessment.
- Slides: 15