Forecasting Brazilian Inflation Using a Large Data Set
The objective of this paper is to verify if exploiting the large data set available to the Central Bank of Brazil, makes it possible to obtain forecast models that are serious competitors to models typically used by the monetary authorities for forecasting inflation. Some empirical issues such as the optimal number of variables to extract the factors are also addressed. I find that the best performance of the data rich models is usually for 6-step ahead forecasts. Furthermore, the factor model with targeted predictors presents the best results among other data-rich approaches, whereas PLS forecasts show a relative poor performance.
Year of publication: |
2010-12
|
---|---|
Authors: | Figueiredo, Francisco Marcos Rodrigues |
Institutions: | Central Bank of Brazil, Research Department |
Saved in:
freely available
Saved in favorites
Similar items by person
-
Evaluating Core Inflation Measures for Brazil
Figueiredo, Francisco Marcos Rodrigues, (2001)
-
A Volatility and Persistence-Based Core Inflation
Filho, Tito Nícias Teixeira da Silva, (2014)
-
Preços Administrados: projeção e repasse cambial
Alves, Paulo Roberto de Sampaio, (2013)
- More ...