Statistical Implications of Inflation Targeting

14 The Experience of Banco Central do Brasil

Carol Carson, Claudia Dziobek, and Charles Enoch
Published Date:
September 2002
  • ShareShare
Show Summary Details
Gustavo Bussinger

Brazil introduced an inflation-targeting regime in mid-1999. The new policy framework was adopted after signs of declining inflation became clear. In January, following the floating of the currency, the exchange rate depreciated significantly. The base interest rate was raised to a high level to curb the impact of the depreciated currency on inflation. Three-year-forward targets within a tolerance interval of plus or minus 2 percentage points for the headline consumer price index provided the nominal anchor for monetary policy. Inflation targeting in Brazil has been implemented along the lines discussed by Bogdanski and others (2000).

Mishkin and Savastano (2000) define inflation targeting as a strategy characterized by “an institutional commitment to price stability as the primary goal of monetary policy, to which other goals are subordinated.” For the central bank to achieve the targets set for inflation, it needs to make use of “any information that is relevant to the forecasting of inflation” (Bernanke and others, 1999). Bernanke and others indicate that inflation targeting implies the use of more information than intermediate targeting strategies.1 This argument is supported by claims that (1) overreliance on an external forecast could lead to circularity, as agents find it optimal to release forecasts of inflation equal to the targeted value, and (2) tying monetary policy too strongly to forecasts of future inflation could, under certain conditions, lead to multiple equilibria. Bernanke and Woodford (1997) raise the point that the central bank should make use of a full structural model capable of analyzing the effects of various policy paths. Mishkin and Savastano (2000) define inflation targeting as an “information inclusive strategy in which many variables, and not just monetary aggregates or the exchange rate, are used for deciding the setting of policy instruments.”

The aim of this chapter is to assess the use of information required to operate an inflation-targeting regime. The first section discusses the use of information in inflation-targeting regimes. Information is used to run models and analyze policy alternatives, but it is also used to communicate policy objectives and intentions to the public.

The second section reviews the implementation of inflation targeting in Brazil, including a review of the information set and the models used by Banco Central do Brasil. The section also discusses how the minutes of the Monetary Policy Committee and inflation reports are used to communicate with the public. The three dimensions of the communication strategy are discussed: openness, transparency, and timeliness.

Section three reviews the key features of the two monetary policy regimes in Brazil: before 1999, when monetary policy was based on administering a pegged exchange rate system, and after 1999, when Brazil adopted an inflation-targeting regime.

The fourth section analyzes the content of the minutes of the monthly meetings of the Monetary Policy Committee since it was created in 1997. The focus of the analysis is to detect whether the use of information to run monetary policy has changed over the two different regimes.

The conclusion indicates that most of the additional resources used to operate an inflation-targeting regime in Brazil have been applied to handling information.2 Before 1999, in the period of the pegged exchange rate system, information was used mainly in the analysis of basic statistics related to past performance of the economy and inflation. Under inflation targeting, the set of data and statistics used has remained roughly the same. But the handling of information has changed in two visible directions. Information is used to build a forward-looking quantitative assessment of the economy and inflation through structural models. This includes the modeling of inflation expectations using data gathered through regular surveys of market expectations carried out by the central bank. Besides that, information is used as a communication tool to facilitate the public’s understanding of policy objectives and intentions.

The Information Content of Inflation-Targeting Regimes

The handling of information in inflation-targeting regimes goes beyond the need for data to feed econometric models that forecast inflation and simulate the impact of alternative policies. Because targets for inflation may be better understood by the public than monetary aggregates or other nominal targets, providing good information to the public in a transparent manner is essential. Also, the costs associated with inflation variability can be reduced through effective communication. Effective communication enhances credibility because keeping the public well informed may increase the likelihood that the central bank will reach the target. Even when the target is missed, explaining the reasons for not reaching it helps to minimize any loss of credibility.

The public’s understanding of the projections is just as important as the clear understanding of how projections are made. Thus almost all inflation-targeting central banks periodically release well-crafted reports on monetary policy and inflation. In general, these reports contain extensive explanations of both the models and their projections in language that is accessible to the general public. Inflation reports are read by an audience wider than economists and policy specialists. Because expectations are an important component of inflation projections, reports are clearly expectations builders.

Inflation Targeting in Brazil: Information Requirement and Transparency Issues3

On July 1, 1999, Brazil formally adopted inflation targeting as the monetary policy framework, following Presidential Decree No. 3088 of June 21, 1999, and the National Monetary Council’s Resolution No. 2615 of June 30, 1999. The target for inflation, verified in December each year, is the 12-month accumulated headline broad consumer price index, calculated by the National Bureau of Geography and Statistics.

Monetary policy decisions are taken by the Monetary Policy Committee, which is chaired by the central bank governor and includes seven vice presidents as voting members and heads of selected central bank departments as nonvoting members. The committee gathers once a month to decide on the interest rate. Decisions are made by simple majority and are based on a wide array of available information.

A set of models with various specifications are reviewed, and their projections and associated probability distributions are analyzed by the committee. The main models are as follows:

  • macroeconomic model of the Brazilian economy (under development);
  • small structural model of the Brazilian economy;
  • core inflation models;
  • leading indicator models;
  • vector autoregressive and Bayesian vector autoregressive models;
  • market expectations surveys;4 and
  • other relevant statistics and economic and financial series.

Transparency is perceived as a key issue for the inflation-targeting policy framework in Brazil. It involves openness as well as an effective communication strategy. Openness means allowing people to monitor the decisions of the committee, and effective communication is related to people’s ability to understand the committee’s decisions, including the supporting evidence. Effective communication also involves comprehensiveness and timeliness.

Comprehensiveness means the public will have all the information required to understand and replicate experiments. In Brazil, the central bank ensures comprehensiveness with a variety of instruments:

  • Minutes of committee meetings are released eight days after meetings. This time lag was lowered from two weeks in 2000.
  • Quarterly inflation reports discuss “the main issues related to the performance of the inflation-targeting regime in Brazil. They include detailed explanations of the results delivered by past decisions and a prospective analysis for future inflation, with special emphasis on the assumptions made in the forecasting process that generated the monetary instrument decisions. Minutes of the previous Monetary Policy Committee meetings are republished in the reports” (Bogdanski and others, 2000).
  • A working paper series has as its main objective the publication of technical work carried out by the staff of the central bank, in all its relevant areas. The series also accepts the submission of external technical papers by authors who have institutional relationships with the Bank. The research department coordinates the process of submission, revision, and publication of the technical work. Revision follows a double-blind process. Working papers can be viewed on the central bank’s website,
  • Technical notes are an additional channel for technical information about the operations of the central bank. The main purpose of the notes is to disclose to the public preliminary results of economic research, as well as the methodologies used to analyze the data released by the central bank. The economics department is in charge of coordinating evaluation and revision of the notes. The vice president of economic policy issues the final clearance for publication. Technical notes of the central bank are available on the Bank’s website also.
  • The central bank carries out a very active, regular program of seminars by external researchers, workshops, and visiting scholars.

Information Requirements and Monetary Policy Regimes

Before 1997, the central bank did not have sole responsibility for conducting monetary policy. Decisions were taken by the Bank’s Collegiate Directorate in both ordinary and irregular meetings. They were announced to the public, but no formal records were kept of the motives driving the decisions. In monthly bulletins and annual reports, the focus was on past performance of inflation and ex post analysis of its behavior.

In January 1998 the Monetary Policy Committee (COPOM) began meeting regularly, having been created as an organized forum to make monetary policy decisions. Its membership consists of central bank directors. Although COPOM was created by a low-level legal instrument, it enjoyed a high degree of independence on monetary policy decisions. The minutes of the meetings began to be disclosed, providing background information about discussions and the rationale for decisions, but there still were long lags between the meetings and the disclosure.

In 1999 Brazil adopted inflation targeting. Clear responsibilities were assigned to the central bank for reaching three-year-forward targets for inflation. The central bank governor is held accountable for monetary policy performance. The legal level of COPOM’s mandate was elevated. Minutes of the meetings began to be disclosed on a timely basis. The minutes contain extensive evaluation of inflation trends and the hypotheses used in their assessment. Quarterly inflation reports contain numerical projections for inflation and the distribution of outcomes.

A review of the 67 meetings of the committee, as reported in the published minutes, indicates what information the members analyzed in support of their decisions. It also provides a case study of the information requirements of an inflation-targeting regime. Specifically, it shows how information requirements have changed since the shift in the monetary policy regime in early 1999.

The minutes are public documents that are made available on the central bank’s website. They are a rich source for the information presented to COPOM members when they gather each month to decide on the course of monetary policy. Minutes are now equally available for the two periods when different monetary policy regimes were in force: 1998 and the beginning of 1999, when Brazil followed an exchange rate-targeting monetary policy; and 1999 and beyond, when Brazil adopted an inflation-targeting framework to guide monetary policy. Analysis of the content of the minutes indicates the changes in focus over the two regimes. The differences are mainly two:

  • The way supporting evidence for decisions is organized and presented has changed.
  • Structural economic models, which were mainly absent in early presentations, began to be employed by the committee in mid-1999. Structural models have been widely used as guides to discussion. Increasingly, market participants are taking an interest in the structure of the models and their results. The use of model information as a communication tool responds to the public’s demand for increased transparency and accountability. A full section describing the key assumptions used in decision making and the prospects for inflation has been included.

The Content of the Minutes in Two Monetary Policy Regimes

This section assesses the need for information under inflation-targeting regimes by comparing the minutes of the meetings in the period up to January 1999, when monetary policy was governed by a pegged system, with those from the meetings held after mid-1999, under an inflation-targeting regime.

The Monetary Policy Committee of Banco Central do Brasil gathered for the first time on January 28, 1998. The minutes from the first meeting show that members assessed the following topics: (1) monetary aggregates and credit, (2) public finances, (3) prices and the level of activity (4) the balance of payments, (5) the external environment and the evolution of the foreign exchange market, (6) bank liquidity, (7) monetary market and open market operations, and (8) monetary policy directives. The structure and the topics discussed remained essentially unaltered over the first period.

In the second period, which began in July 1999 after the formal adoption of inflation targeting, the committee’s discussions covered all of the topics above. But there is an important change. A full section on prospective evaluation of inflation trends was added, reflecting a significant change in terms of incorporating a forward-looking approach to guide monetary policy. In this period, the minutes indicate that committee members reviewed inflation models presented to them, the hypothesis used, and the results that were obtained.


The minutes of the Monetary Policy Committee meetings are a rich source of documentation on the information set used by its members in making their decisions.

A review of the public documents shows that the minutes of the first period are longer, descriptive, and backward looking. Structural models used to forecast inflation are absent. In the second period, under inflation targeting, a quantitative forward-looking assessment of inflation based on structural models was introduced. The structure of the models is disclosed so results can be replicated. The minutes indicate that model results along with data on expectations and other relevant information have been important inputs in guiding policy decisions.

In the case of Brazil, the handling of information has changed. Before 1999, very large amounts of data were collected and discussed. The inflation-targeting period improved modeling. With the exception of expectations data, almost all data series and statistics had already been used. Most of the additional resources used in the handling of information under the inflation-targeting regime in Brazil have been applied to developing models designed to forecast inflation and to simulate the impact of policy alternatives.


A related issue presented by the authors, not dealt with here, is that the implementation of inflation targeting would be more difficult than other policy strategies that make extensive use of intermediate targets—such as monetary aggregates or the exchange rate—because presumably the central bank is able to control such intermediate targets with a high degree of precision.


The handling of information is defined as the way relevant information is organized, processed, and disclosed.


This history of implementing inflation targeting in Brazil is based on Bogdanski and others (2000).


Market expectations surveys are carried out by the central bank with around 70 banks and economic consultants on a high-frequency basis. Summary statistics of the surveys are disclosed every week, along with a rank of the top five institutions with the most accurate forecast for key macroeconomic variables. The fact that monthly rankings identify institutions by name functions as an additional incentive for participants to put their best efforts into responding to the surveys.

    Other Resources Citing This Publication