Brazil: Selected Issues


Brazil: Selected Issues

Upgrading Brazil’s Inflation-Targeting Framework1

Brazil was one of the early adopters of an inflation target in 1999. Its framework has proved resilient and inflation has stabilized at significantly lower rates than in previous decades. Since then, lessons from international research and the experience from other inflation-targeting countries have shown that it is possible to achieve better inflation and growth outcomes by enhancing central bank credibility and transparency. Brazil’s experience also suggests that there could be room for improvement, as both inflation and inflation expectations have often deviated from the mid-point of the target range, increasing the cost of disinflation. This paper discusses how Brazil could upgrade its inflation-targeting framework in terms of its design, analytical framework and effectiveness of central bank communication. Central bank independence is the top priority, but Brazil could also benefit from adopting an inflation-forecast targeting framework. Whereas this paper focuses on monetary policy, a systematically consistent policy mix is also important to secure better inflation and growth outcomes.

A. Why Upgrade Brazil’s Inflation-Targeting Framework?

1. Brazil’s inflation-targeting framework has proved resilient, but inflation has often deviated from the mid-point of the target range. Brazil was one of the early adopters of an inflation-targeting (IT) framework in 1999 as the exit from a pegged exchange rate regime created the need for a new nominal anchor. The framework has proved resilient and inflation has stabilized at similar rates to those observed in the years immediately after the implementation of the Plano Real (1995−98) and significantly below those experienced during hyperinflation in previous decades.2 However, inflation has often deviated from the mid-point of target range. Between 1999 and 2015, annual IPCA inflation oscillated between 3 percent and 13 percent, whereas the mid-point of the target remained close to 4.5 percent. In seventeen years of inflation-targeting, annual IPCA inflation was above the mid-point of the target range or its upper limit in 13 and 4 years, respectively. It has never been below the lower limit of the target range (Figure 1).

Figure 1.
Figure 1.

Brazil Annual CPI Inflation, 1995-2015

(In percent)

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

Source: Brazilian Institute of Geography and Statistics (IBGE) and Banco Central do Brasil.

2. Reducing inflation and inflation volatility is important for a number of reasons.3 High inflation is often associated with high inflation volatility.4 This has been the case in Brazil since it adopted an IT system, as inflation volatility has come mostly from inflation outcomes being above the target (Figure 1). High and volatile inflation is not only costly to reduce, but it can also impair the efficient allocation of resources across the economy. Uncertainty about inflation also makes it more difficult for businesses and households to plan and make spending and saving decisions. Importantly, high inflation disproportionately affects low-income households, who have relatively limited access to financial instruments to protect their nominal income against inflation.

3. International experience suggests that it is possible to achieve better inflation and growth outcomes. IT became a popular monetary policy framework since it was first implemented by New Zealand in 1989.5 It has now been adopted by about 30 central banks around the world. The framework has evolved in recent decades as new research and practice have shed light on how to achieve better outcomes, particularly in terms of lower volatility of inflation and GDP growth. By upgrading its IT framework, Brazil might also benefit from better economic outcomes. Indeed, many inflation-targeting countries (including in Latin America) have experienced lower inflation and output volatility than Brazil since 1999 (Figure 2).6 Of course, the divergence in output and inflation volatility across countries also reflects differences in the structure of the economy and the type and magnitude of the shocks it is subject to. But the sample includes both advanced economies—which have been severely affected by the global financial and the European debt crises—and other emerging market economies—which, like Brazil, are particularly vulnerable to additional shocks, such as commodity price fluctuations and swings in capital flows. So, taken together, these also suggest that it might be worth investigating whether Brazil has room for enhancing its monetary policy framework.

Figure 2.
Figure 2.

Standard Deviation of Inflation and Output Across IT Countries, 1999-2015 (1)

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

Source: Fund Staff calculations.(1) The sample includes only the years in which inflation-targeting was in place and countries where inflation-targeting had been introduced by the early 2000s, namely Australia, Brazil, Canada, Chile, Colombia, Czech Republic, Hungary, Iceland, Israel, Mexico, New Zealand, Norway, Peru, Philippines, Poland, South Africa, South Korea, Sweden, United Kingdom and Thailand.(2) In percentage points. Target has changed over time in some countries.(3) Output gap defined as the percent deviation of output from potential output. Based on central banks estimates where available and on HP filter estimates otherwise.

4. Lessons from theory and practice in recent decades have highlighted the role of expectations in achieving better inflation and growth outcomes. The key insights are that monetary policy is more effective when: (i) the central bank is credible such that inflation expectations are anchored; and (ii) when monetary policy can influence expectations of future interest rates, which the central bank can achieve with enhanced transparency. Over time, many central banks have adopted an inflation-forecast targeting framework, which has credibility and transparency as core principles. These are the topics covered in the next sections. Section B discusses why managing expectations (through improving central bank credibility and transparency) is crucial for increasing the effectiveness of monetary policy and how inflation-forecast targeting incorporates that insight. Section C discusses how Brazil can upgrade its IT framework to achieve lower inflation and output volatility. Section D uses a simple model to illustrate the potential benefits for Brazil of upgrading its IT framework.

5. Whereas this paper focuses on monetary policy, a systematically consistent policy mix is also important to secure better inflation and growth outcomes. Adopting measures to enhance the transmission of monetary policy and to promote broader macro stability are also important to achieve lower output and inflation volatility. For example, the large share of subsidized credit extended by public banks weakens the effectiveness of monetary policy, requiring higher policy interest rates to achieve a given inflation outcome.7 And widespread price- and wage-indexation practices augment the effect of price shocks and spread their impact across the economy. In particular, the minimum wage indexation mechanism has contributed to the erosion of competitiveness, as real income gains in excess of productivity growth have boosted unit labor costs.8 In addition, sound fiscal and macroprudential policies are essential for price stability. A deteriorating fiscal position and concerns over fiscal dominance pose challenges to the conduct of monetary policy.

B. Credibility, Transparency and Inflation-Forecast Targeting

6. Modern academic research and international best-practice have shown that effective monetary policy relies on successfully managing expectations. There are two key aspects to this finding: (i) the role of central bank credibility and anchored long-run inflation expectations; and (ii) the role of transparency in central bank communication in influencing expected future short-term interest rates and in anchoring inflation expectations. Both of these aspects are core principles of inflation-forecast targeting (IFT).

How can central bank credibility reduce inflation and output volatility?

7. Reducing inflation volatility often comes at a cost of higher output volatility in the short run. Economies are often hit by shocks that move inflation and output in opposite directions. For example, higher energy prices push up on inflation, but reduce real disposable income and thus weigh on demand and growth. By trying to bring inflation back to target in response to such shocks, monetary policy would need to exacerbate the deviation of output from potential in the short run: in trying to reduce inflation volatility, monetary policy increases output volatility.9 The policy decision involves choosing the best trade-off between output and inflation volatility in the short run, while anchoring long-term inflation expectations to the target.10

8. Central bank credibility can reduce both inflation and output volatility in the short run. Despite this short-run trade-off, monetary policy can achieve a better balance between inflation and output volatility when the central bank is credible.11 This is because anchored inflation expectations reduce inflation persistence and increase the efficacy of monetary policy, reducing both the impact on inflation of a given shock and the change in output needed to bring inflation back to target. When inflation expectations are not anchored, shocks that move inflation away from the target will have an extra effect on future inflation through two main channels:

  • a. Price and wage setting behavior. By pushing up wages and prices, higher inflation expectations could lead to inflation becoming more persistent. If households expected inflation to remain high, they might demand higher nominal wages in order to avoid losing purchasing power. If firms expected prices to rise in the future, they might raise the prices of the goods and services they produce, and might also choose to raise wages. Recent survey evidence suggests that current inflation is a key driver of pricing decisions in Brazil.12 The impact of current inflation on future prices and wages can be larger in the presence of price and wage indexation, as it is the case in Brazil. Indexation automatically increases inflation persistence and is itself a symptom of low monetary policy credibility.

  • b. Expected real interest rates, exchange rates and asset prices. Inflation expectations act as shock amplifiers or absorbers depending on whether the central bank is credible. When the central bank is not credible, a shock that shifts inflation above the target will raise inflation expectations, lowering expected real interest rates, boosting asset prices (through the lower real discount rate) and depreciating the exchange rate. These would affect households and companies spending decisions and stimulate demand, amplifying the upward pressure on inflation. In contrast, if the central bank were credible and committed to adjusting policy to bring inflation back to target following the adverse shock, inflation expectations would remain anchored, current and future short-term real interest rates would be higher, resulting in an exchange rate appreciation and lower asset prices, which would weigh on demand and help absorb the impact of the original shock.

As a result, when inflation expectations are anchored, both policy interest rates and output need to move by less to bring inflation back to target. So whereas a central bank will always face a short-term trade-off between stabilizing inflation and output, this trade-off will be better when monetary policy is credible, resulting in lower inflation and output volatility over the medium-term.13

How can transparent central bank communication reduce inflation and output volatility?

9. Effective monetary policy relies on successfully influencing expectations about the future path of short-term interest rates. This is because whereas the central bank controls the short-term policy interest rate, what matters most for private-sector spending decisions (and thus inflation and output) is the impact of monetary policy on the public’s expectations about future policy and, as a result, on longer-term interest rates, asset prices and the exchange rate.14

10. Transparent central bank communication is crucial to influence expectations of future policy rates and to secure credibility.15 Policy decisions that are neither well-communicated nor clearly explained could have little or even the opposite effect on expectations of future policy rates, and thus on the economy. They could also damage central bank credibility and de-anchor inflation expectations. In contrast, the more the public understands the aims of future policy and how the central bank would react in various circumstances, the less monetary policy itself might need to adjust. And the more the central bank explains its reasoning in detail, the more convincing the promise to return inflation to target becomes. Improving transparency through communication about what variables the central bank responds to, how aggressively and why is crucial to improve policy effectiveness, and thus achieve lower inflation and output volatility.16

The policy frontier: inflation-forecast targeting

11. Credibility and transparency are core principles of inflation-forecast targeting. As the critical role of credibility and transparency has become clearer, both advanced and emerging market economies have upgraded their frameworks to flexible inflation-targeting, or inflation-forecast targeting (IFT). Nearly two thirds of the countries currently with an inflation target (or objective) are inflation-forecast targeters. The three distinguishing features of IFT central banks—which embed the role of credibility and transparency—are:

  • I. A well-defined long-term target

  • II. Forecasts based on endogenous interest rates and exchange rates

  • III. Central bank’s inflation forecast used as intermediate target and communication tool

In practice, IFT frameworks function as follows. First, a well-defined long-term target (namely a continuous point target) is announced. Second, the central bank publishes a forecast where inflation returns to the target over its preferred horizon and explains which policies are consistent with this forecast. Finally, it credibly communicates the rationale behind the policy decision, as well as why previous forecasts were or were not fulfilled and the key judgments and risks around the latest forecasts. IFT frameworks also share a number of features with traditional IT frameworks, namely central bank independence; the use of the inflation target as a nominal anchor; accountability mechanisms; and the need for sound fiscal and macroprudential policies to support monetary policy. 17

12. Well-defined long-term targets increase the effectiveness of monetary policy.18 Two characteristics of an inflation target are essential for anchoring inflation expectations and thus improving the effectiveness of monetary policy. First, clarity about what the central bank’s target truly is, allowing long-run inflation expectations to anchor around it. Second, the target should acknowledge both the short-run output-inflation trade-off and the long lags in the transmission of monetary policy. Effectively, these mean that it is neither possible (given lags) nor optimal (given the trade-off) to keep inflation on target at all times or to aim to hit the target at a given time (eg: at the end of the year). An inflation target defined as a range to be reached at the end of the year, as it is the case in Brazil, is at odds with these concepts. The target range provides flexibility, which can be used to accommodate shocks that monetary policy cannot or chooses not to offset by the end of the calendar year. But this comes at the cost of lack of clarity about what inflation rate the policymaker truly wants. This can be particularly challenging when the central bank is not independent. In periods of persistent deviation of inflation from the target, this could lead inflation expectations to anchor at the upper or lower bounds of the target range instead of its mid-point. Also, because of the lags with which monetary policy operates, occasionally it might be optimal to allow year-end inflation to be outside the target range (for example, when large price level shocks happen close to the end of the year). But that could also damage credibility. In contrast, a well-defined long-term target, namely a continuous point target, is clear about what inflation rate the policymaker truly wants. Its long-run focus communicates that whereas inflation will remain close to the target on average,19 it will deviate from it when optimal, provided long-term expectations remain anchored and the central bank credibly communicates how it will bring inflation back to target (Figure 3 and Table 1).20

Figure 3.
Figure 3.

Illustrative Representation of Target Definitions

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

Table 1.

Inflation Target Credibility and Transparency Checklist

article image

13. Forecasts based on endogenous interest and exchange rates enhance transparency, credibility and thus make monetary policy more effective. Under IFT, the monetary policy committee (MPC) chooses, at each policy meeting, a policy-rate path consistent with the forecast for inflation being at target over the forecast horizon. This approach is superior to producing forecasts based on exogenous interest rates and exchange rates (such as constant or market-implied paths). This is because it recognizes that it is the entire path of expected future policy rates (rather than the MPC’s current target for the short-term policy rate) that affects the interest rates which are relevant for output and inflation. The approach also allows the central bank to influence the expected path of future short-term rates, making monetary policy more effective.21 Some IFT central banks influence expectations by describing their preferred interest rate path with words. The most transparent—such as in the Czech Republic, New Zealand, Norway and Sweden—go as far as publishing that path. The Federal Reserve Board has moved in a similar direction, by publishing FOMC participants’ median assessment of the appropriate path for policy rates. Importantly, publishing the MPC’s current preferred path for interest rates does not preclude it from deviating from it when the future comes. The fear that the public would understand these paths as a promise has proved unwarranted by the experience of the countries who have published it.

14. Inflation forecasts in IFT central banks are effectively an intermediate target and a key communication tool. As it is neither possible nor ideal to keep inflation at target at all times, the inflation forecast based on endogenous interest rates is effectively an intermediate target, as it communicates at what speed the central bank plans to bring inflation back to target. It incorporates the preferences of the policy makers, their views about the economy and about the transmission mechanism of monetary policy. In addition to using the forecast as a communication tool, IFT central banks aim to convincingly explain the reasoning behind individual policy decisions; publish their assessment of the economic outlook as well as the analysis behind that assessment; and discuss the risks around those views. Taken together, these allow IFT central banks to be transparent about how they expect to manage the short-run output-inflation trade-off. This, in turn, helps affect both inflation expectations and expectations of future policy, making monetary policy more effective.

C. How can Brazil Achieve a Better Short-run Output-Inflation Trade-off?

15. Brazil could benefit from upgrading its IT framework in terms of its design, analytical framework and effectiveness of central bank communication. The purpose of this section is two-fold. First, it discusses evidence on central bank credibility and transparency in Brazil. Second, it presents key measures Brazil could take to upgrade its IT framework. Over time, these measures would increase the credibility and effectiveness of monetary policy by affecting the public’s expectations of future policy and anchoring inflation expectations, allowing for better inflation and output outcomes.

C.1 Central bank credibility and transparency in Brazil

16. Credibility is typically measured by comparing inflation expectations to the target over the medium– to long-run. Whereas central bank credibility is not directly observable, better- anchored IT frameworks tend to have smaller deviations of inflation expectations from the target, lower dispersion of inflation forecasts, and minimum correlation between long-run inflation expectations and current inflation or inflation surprises. Focusing on longer horizons (typically beyond 5-years-ahead, but at least 3-years ahead) is important because short-term measures are highly correlated with current inflation – which is affected by price level shocks – and because it accounts for the long lags of monetary policy transmission. Another way of assessing credibility is to compare private-sector forecasts with those of the central bank (typically available at shorter horizons). Finally, it is important to capture a representative set of expectations: international best-practice has been to regularly monitor expectations from a wide set of economic agents, including households, firms, professional economic forecasters and financial markets.22

17. Available data suggest that Brazil can improve central bank credibility. Whereas data on inflation expectations are relatively limited for Brazil, available information suggests the system could be better-anchored.23 In Brazil, the deviation of longer-term inflation expectations from the mid-point of the target range has been higher than in other countries, particularly the most anchored inflation-forecast targeters, such as Canada and Chile (Figure 4).24 Large deviations have been frequent in recent years (Figure 5). At the same time, market participants have consistently expected higher inflation than the central bank’s forecast produced using interest rate expectations from market participants (Figure 6). The dispersion of inflation forecasts in recent years has also been somewhat high relative to the experience of the top countries (Figure 7). Finally, previous Fund Staff work has shown that, starting in 2011, upside inflation surprises had begun to affect inflation expectations at 18-month horizons.25 Taken together, these suggest that there is room for improving central bank credibility in Brazil.

Figure 4.
Figure 4.

Longer-Term Inflation Expectations, 1999-2015 Average (1)

(Absolute deviation from target, percentage points)

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

Source: Consensus Forecasts and Fund Staff calculations(1) 3-year-ahead inflation expectations(2) Average of Colombia, Mexico and Peru from start of respective inflation-targeting framework
Figure 5.
Figure 5.

Longer-Term Inflation Expectations in Brazil (1)


Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

Source: Consensus Forecasts and Fund Staff calculations(1) 3-year-ahead inflation expectations
Figure 6.
Figure 6.

Short-Term Inflation Forecasts (1-year-ahead)

(In Percent)

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

Source: Banco Central do Brasil.
Figure 7.
Figure 7.

Dispersion of Long-Term Inflation Expectations, 2011-2015 Average(1)

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

Source: Consensus Economics and Fund Staff calculations(1) 3-year-ahead inflation expectations, standard deviation of sample of forecasts(2) Average of Colombia, Mexico and Peru

18. Brazil could also make monetary policy more effective by increasing central bank transparency. Both research and international practice have shown that increased transparency in central bank communication makes monetary policy more effective (Section B). As such, central bank transparency has increased significantly across the globe in recent years. It also played a key role during the Great Financial Crisis, helping advanced economies overcome challenges posed by the effective lower bound for policy interest rates. Whereas measuring transparency is inherently difficult, published academic research has relied on transparency indices such as the one produced by Dincer and Eichengreen, which scores central banks in a number of categories.26 According to this comprehensive measure, central bank transparency in Brazil has tended to be higher than in other Latin American IFT countries. Brazil since the IT framework was adopted, whereas it has been steadily increasing in other central banks, particularly those currently at the frontier. The gap between the Banco Central do Brasil (BCB) and the most transparent central banks is large and has widened, suggesting there is room for improvement (Figure 8).

Figure 8.
Figure 8.

Dincer and Eichengreen Central Bank Transparency Index

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

(1) Average of Chile, Colombia, Mexico and Peru(2) Average of all central banks with an inflation targetSource: Dincer and Eichengreen (2014)

C.2 What could Brazil do to upgrade its inflation-targeting framework?

19. Brazil could reduce output and inflation volatility by increasing central bank independence and upgrading to an inflation-forecast targeting framework. This section discusses specific measures, summarized in Table 2.

Table 2.

Credibility and Transparency Best-Practice Checklist

article image
Central Bank Independence

20. The case for central bank independence is well-established in the literature and has been validated by international experience. The technical argument in favor of central bank independence is simple: an independent central bank is more credible because it is free from either perceived or actual political interference to either boost short–run growth or finance the budget deficit at the expense of higher inflation. Increased credibility translates into better-anchored inflation expectations, allowing the central bank to achieve better inflation and output outcomes. This has been well-established in an extensive literature which dates back to the late 1970s and continues to be supported by today’s policymakers.27 Also, international experience shows that central bank independence has been associated with better inflation outcomes with no cost to economic growth.28

21. Brazil is an outlier for its lack of ‘legal’ (or ‘de jure’) central bank independence.29 The two distinctive features which Brazil lacks are: (i) a law that gives the central bank the mandate to achieve monetary and financial stability,30 and (ii) fixed-term appointments for members of the monetary policy committee (MPC) which do not overlap with the political cycle (to insulate them from political pressure) (Figure 9). Once de jure independence is granted, de facto independence is secured through systematically appointing analytically strong board and staff members. Central bank independence should be accompanied by mechanisms to ensure the central bank remains accountable to the public and it has enough resources to achieve its mandates. These mechanisms would be part of the central bank law/act and often take the form of regular testimonies to Congress by the central bank governor (which the BCB already has in place) and members of the MPC, as well as annual reports on the central bank’s management and financial accounts.

Figure 9.
Figure 9.

Terms of Employment of MPC Members: Length of Contract (1)

(In years)

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

(1) Average of fixed contract legth for Governor and Deputy GovernorsSource: Hammond (2012)
A well-defined long-term target

22. Brazil could better anchor inflation expectations and improve central bank credibility by adopting a well-defined long-term target. Brazil’s design of its inflation target follows international best-practice in that it uses as benchmark a reliable index of headline inflation which is not computed by the central bank itself. But there is room for improvement. The target is defined as a ‘tolerance interval’ which allows inflation to range between 2.5 percent and 6.5 percent and is considered met as long as inflation is anywhere within this interval at the end of the year.31 This is in contrast to IFT countries, which have long-term point targets (Figure 10).32 The distinction between a point and a range target, as well as the difference between a short- and long-term horizon are both important. The lack of a well-defined long-term point target permits outcomes which, while formally in compliance with the framework, can erode its credibility and contribute to a de-anchoring of long-run inflation expectations (Section B).

Figure 10.
Figure 10.

Inflation Targets, Cross-Country Comparison(1)

(In percent)

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

(1) Established IFT central banks and BrazilSource: Central banks or monetary authorities
A structured forecasting framework with endogenous interest rates and exchange rates

23. The BCB could increase policy effectiveness and credibility by enhancing its forecasting and policy analysis framework. Through an IFT forecasting and policy analysis system (FPAS), BCB staff would provide economic analysis and produce a model-based staff forecast using endogenous interest rates and exchange rates. The staff forecast would be used as a baseline for the MPC’s discussions of the economic outlook, policy options and scenario analysis. The MPC forecast would then incorporate the MPC’s view about the economy, their preferred path for policy rates as well as their assessment of the balance of risks.33

Effectiveness of central bank communication

24. The BCB uses most of the central bank communication tools adopted internationally. For example, it publishes a quarterly Inflation Report with updated inflation and GDP forecasts and it holds a press conference to explain those. COPOM’s policy decisions are accompanied by press statements and minutes from the policy meeting and the votes of the committee members are published. Also, the background material presented to the COPOM is released following a deferral period. Finally, open letters and testimonies to Congress are used as accountability tools. One area where the BCB could expand its toolkit is through a more extensive use of speeches and media interviews by COPOM members.34

25. The room for transparency improvement in the case of the BCB is mostly on the effectiveness of its communication rather than on the range of tools used. In particular, official central bank communication could benefit from: a) publishing endogenous inflation and growth forecasts, ie: forecasts which are internally consistent with the COPOM’s preferred policy path (even if the policy path itself is not published), as opposed to based on exogenous interest and exchange rates; b) placing less emphasis on a factual description of historical data in favor of a more in-depth assessment of the key drivers of growth and inflation;35 c) explaining how (and why) these key drivers are assumed to evolve in the central forecast (and, when updating the forecast, discussing how these evolved relative to expectations); d) better explaining monetary policy strategy issues and the motivation behind the policy decision (namely, policymakers’ preferences on how to manage the short-term output-inflation trade-off); and e) providing a more substantive discussion of the risks around the central forecast.

26. The latest changes to the COPOM’s communication style are an important step in the right direction, but further progress can be made. In recent months, a number of important changes have been introduced by the BCB, affecting the Inflation Report, the policy statement and the minutes from the policy meeting (now called ‘notes’). These have helped improve central bank communication. In particular, there is less emphasis on a factual description of historical data in favor of a more detailed discussion of the balance of risks and the divergence of views within the committee. But there is scope for further gains from introducing new enhancements, as discussed above and summarized in Table 2.

Central Bank governance

27. Measures to enhance MPC credibility and internal transparency could also be beneficial. Whereas central bank governance is a broad topic, we focus on three measures which could be particularly helpful for credibility and transparency:

  • a. First, it is best-practice for the MPC to follow a code of conduct for external engagement. Such code would provide guidelines for MPC members on issues such as liaison with the media, speeches, political involvement, conflicts of interest and purdah (or blackout) periods for external engagement, amongst others.36 Such guidelines would serve to reinforce the reputation and integrity of the monetary policy process.

  • b. Second, it is also best-practice for the MPC to be composed only by individuals with expertise in the field of economics and monetary policy. In Brazil, instead, directors of departments with responsibilities in other fields (such as management, supervision, regulation and external liaison) are also part of the COPOM.37

  • c. Third, increased internal transparency could help strengthen the policy debate. Allowing staff from the monetary policy directorate to follow the policy decision process allows specialists to increase the effectiveness of their analysis, as well as challenge and surface alternative views about the key issues and risks discussed by the MPC. This could be done, for example, by allowing staff to watch meetings at early stages of the policy cycle (such as briefings on the economic outlook, the baseline forecast and selected key forecast issues) and adopting a process whereby senior management debrief staff on the content of more sensitive (eg: policy decision and final forecast) meetings. It is not clear whether this is current practice at the BCB.

Summary table of how Brazil could upgrade its inflation-targeting framework

28. The measures proposed to upgrade Brazil’s inflation-targeting framework should be treated as complements, but there is a sense of priority. Table 2 summarizes key aspects of Brazil’s IT framework which could be upgraded. It is important to treat these measures as complements and not substitutes, as many of them are interlinked. Even so, there is a sense of priority. The lack of complete central bank independence is the most pressing issue, as either actual or perceived government influence on monetary policy could jeopardize the benefits from other measures. Once the central bank is independent, enhancing the definition of the inflation target as well as moving to endogenous policy variables in the forecasting process would facilitate the adoption of measures to improve the effectiveness of central bank communication. However, some of the proposed measures—particularly those concerning governance and effectiveness of communication—could be adopted even before central bank independence. Please see Table 2 for a description of priority and responsibility for each of the proposed measures.

D. Illustrating the Benefits of Upgrading Brazil’s IT Framework

29. This section uses a simple FPAS model to illustrate the benefits of upgrading Brazil’s IT framework. The exercise consists of using an FPAS model to provide illustrative estimates of what the volatility of output, inflation and policy rates would be if Brazil improved its monetary policy framework. The model used is a small open-economy gap model, with equations for output (IS curve), inflation (based on an expectations-augmented Phillips curve), and the short-term interest rate (policy reaction function). The model is forward-looking, in that expectations and the policy reaction function are driven in part by the model’s own future solved values (in the long run, both expectations and outcomes converge to steady-state paths). Demand shocks are represented by the stochastic term in the output gap equation, and supply shocks by that in the headline and core inflation equations. Please see Appendix 1 for more information on the model.

30. The benefits from a better policy framework can be thought of as moving the short-run output-inflation trade-off closer to the frontier, as well as improving the frontier itself. The Taylor Efficiency Frontier (TEF) is defined as the best possible combinations of inflation and output volatility that can be achieved if the MPC adjusted interest rates optimally to stabilize inflation and the output gap, given the demand and supply shocks hitting the economy (Figure 11). Points above the TEF are inefficient as it takes a larger output gap to reduce inflation (Point A). Improving the monetary policy framework could shift the short-term output-inflation trade-off closer to the TEF (Point B), by increasing the forward-looking behavior of households and firms, as well as allowing the central bank to pursue optimal monetary policy. Improving the framework could also shift the frontier inwards (towards lower output and inflation volatility), by reducing the volatility of the shocks hitting the economy. This could be due, for example, to a better anchoring of inflation expectations as monetary policy becomes more credible and the central bank is granted independence.

Figure 11.
Figure 11.

Improving The Short-Run Output-Inflation Trade-Off

Citation: IMF Staff Country Reports 2016, 349; 10.5089/9781475553222.002.A006

31. The FPAS model can provide illustrative estimates of the gains from moving closer to the frontier.38 The results of this exercise are in Table 3, which shows the volatility of selected macro variables in alternative cases. Under a better framework, optimal monetary policy would take into account the entire path of expected short-term interest rates and it would reduce uncertainty about policy objectives and the MPC’s preferences. In the model, this improvement is captured by adopting an optimal control policy, where the policymaker minimizes a loss function which incorporates expected future output gaps and inflation deviations from the target and which also places relatively more weight on large deviations (column 2 vs 1 in Table 3). In addition, a more credible monetary policy would reduce the impact of past price changes on future inflation through reduced inflation indexation. In the model, this can be proxied by a more forward-looking Phillips curve equation. For this illustrative exercise, we assume the degree of backward-looking behavior falls by half (column 3 vs 2). Taken together, a better monetary policy framework would result in lower volatility in output and inflation. The improved framework would actually allow for the nominal policy rate to move less once credibility was established (fourth row), as increased transparency and credibility make monetary policy more effective in managing policy and inflation expectations.

Table 3.

Gains from Adopting an IFT Framework - Illustrative Estimates

(Unconditional Sample and Theoretical Standard Devia

article image
* The baseline unconditional standard deviations are based on the model-based historical reaction function for the BCBSource: IMF Staff estimates

32. These estimates are a lower bound. The illustrative estimates above do not account for the fact that a better framework could also reduce the volatility of the shocks hitting the economy. This means that, by upgrading its IT framework, Brazil would not only move outcomes closer to the frontier, but it could also shift the frontier itself. For Brazil, the main channels would be shifts in confidence and more anchored inflation expectations, which could reduce the volatility of inflation shocks. The gap between Brazil and the top IFT countries in terms of anchoring of inflation expectations (Section C) suggests that the magnitude of the gains through less volatile shocks would likely be nontrivial. And so does international evidence. In the U.K., for example, the introduction of central bank independence and a well-defined point target for inflation in 1997 led to a dramatic reduction in long-term inflation expectations.39 Also, long-run inflation expectations were better-anchored in Canada and Chile (both inflation-forecast targeters) than in the United States before it adopted an explicit numerical inflation objective and started publishing interest rate forecasts in 2012.40

Appendix I. The FPAS Model Equations1

IS Equation

The output gap (y^t) is defined as the difference between the log-level of output (yt) and potential output (y¯t). The IS equation relates Brazil’s output gap (y^t) to past and expected future output gaps, the deviations of the lagged one-year real interest rate and the real effective exchange rate from their equilibrium values, and the rest-of-the-world output gap.


Phillips Curve

In the Phillips curve, core inflation (πtc) depends on inflation expectations (π4t+4c) and past year-on-year core inflation (π4t1c), with coefficients on both terms adding up to one, and the lagged value of the output gap ((y^t1). It also depends on the rate of real effective exchange rate depreciation (Δreer^t), as well as the deviation of the real effective exchange rate from its equilibrium value, as a real depreciation raises the domestic cost of imported intermediate inputs and final goods and creates upward pressure on prices. Finally, we allow some small pass-through from oil and food price inflation to core inflation.


Policy Interest Rate

Monetary policy follows an inflation-forecast-based reaction function.


The three-quarter-ahead inflation projection ((π4t+3Candπ4t+3H) is to ensure more robustness as policy is reacting to a mix of current data, near-term forecast, and model-based projection in the initial periods. Alternatively, monetary policy could follow a risk-management or risk-minimization strategy through a loss function. The loss function incorporates the principal objectives of the central bank in policy making - taking the appropriate actions to bring inflation back to its long-term target over time and closing the output gap.


The loss function or objective function typically associated with inflation targeting uses a quadratic formulation, implying that large deviations are more important in the thinking of central banks than small errors or deviations. The loss function has equal weights in minimizing the deviations of inflation expectations from the inflation target and the output gap. The loss function also includes a term in the change of the policy interest rate. This term has the effect of preventing very sharp movements in the policy interest rate, which could otherwise occur on a regular basis, and which would not reflect the behavior of central banks in practice. By taking account of both current and expected future values of output and inflation, this formulation enables the central bank to incorporate into its decision making any information currently available to it about likely developments in the economy over the next few quarters.

Real Interest Rates and Real Exchange Rates

The real interest rate (rt) is defined as the nominal interest rate minus expected core inflation.


The bilateral real exchange rate between Brazil and the United States (zt) is defined in terms of Brazil’s core CPI (ptC), and in such a way that an increase means a depreciation of the Brazilian real. The equilibrium real exchange rate is assumed to be determined by the equilibrium terms-of-trade.


The real effective exchange rate is the trade-weighted bilateral real exchange rates of Brazil versus seven regions in the world (U.S., Euro Area, Japan, China, Emerging Asia, Latin America, and the rest of the world).


Risk-adjusted UIP Condition

The risk-adjusted uncovered interest parity condition links the bilateral exchange rate between Brazil and the U.S. with the interest rates in the two economies (it and itUS).


The equation allows the expected exchange rate (Est+1) to be a linear combination of the model-consistent solution (st+1) and backward-looking expectations (st-1) adjusted for trend exchange rate depreciation (2/4(Δz¯tπ*US+π*)) The factor ¼ which multiplies the inflation differential (π*,US - π*) de-annualizes the inflation rates (which are expressed in annual terms), while the factor 2 which multiplies the expected trend depreciation (Δz¯t(π*USπ*)/4) is necessary as we extrapolate the nominal exchange rate in the past period (st-1) two-periods into the future using the steady-state growth rate in the nominal exchange rate. The factor 4 before the expected depreciation (Est+1 - st) annualizes the expected quarterly depreciation rate, to make it consistent with the interest rate quoted on the annual basis. A time-varying variable (σrCTRY) is also included to account for shocks to country-risk premium.

Relative Prices

Headline inflation is characterized mainly through the dynamics of relative price movements (core CPI (ptc) relative to headline CPI (ptH)). In the long run, headline inflation is assumed to be equal to underlying (core) inflation, though it can diverge over prolonged periods of time (eg: when there is a trend in the relative prices of non-core items such as unprocessed food or energy). The dynamics of relative prices (rPt) are modeled as the sum of the relative price trend (rp¯t) and the relative price gap (rp^t). The relative price gap depends on the real price of oil and food in international markets adjusted for exchange rate effects, while the relative price trend growth is assumed to be an autoregressive process with mean zero.


Unemployment Rate

The unemployment rate (ut) is characterized by a “gap version” of Okun’s law. The equation implies that a one percentage point increase in the unemployment gap (ût) is associated with approximately two percentage points of negative output gap. The NAIRU (u¯t) is assumed to follow a stochastic process that has both shocks to the level and shocks to the growth rate.


Potential Output

The potential growth rate (Δy¯t) is assumed to converge to its steady state level (Δy¯ss) in the longer term. However, it can deviate from the steady-state level for prolonged periods of time.


Rest of the World

The rest-of-the-world output gap relevant for the Brazilian economy is defined as a weighted average of output gaps in the seven regions (U.S., Euro Area, Japan, China, Emerging Asia, Latin America, and the remaining countries), using export shares as weights.


Commodities and Terms-of-Trade

The real price of oil (rptOil) is defined as the global oil price (ptOil) in U.S. dollars relative to the U.S. CPI (ptUS). In equilibrium, the real price of oil is assumed to grow at a rate of zero, although the actual growth rate can deviate from zero for long periods of time. The real price of oil gap (rp^tOil), defined as the difference between the real price of oil and its equilibrium value, is modeled as an autoregressive process with a shock.


We follow similar modeling strategy for the real price of food.



  • Alesina, A. and L. Summers, 1993, “Central Bank Independence and Macroeconomic Performance: Some Comparative Evidence” in Journal of Money, Credit and Banking, Vol. 25, May, pp. 151162.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Alves, S. A. L., 2014, “Lack of Divine Coincidence in New Keynesian Models”, Journal of Monetary Economics, Vol. 67, July, pp. 3346.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Anderson, G and B. Maule, 2014, “Assessing the Risk to from Inflation Expectations,” Bank of England Quarterly Bulletin, Vol. 54, No. 2, pp. 14862.

    • Search Google Scholar
    • Export Citation
  • Barro, R. J., and David B. Gordon, 1983, “Rules, Discretion and Reputation in a Model of Monetary Policy,” Journal of Monetary Economics, Vol. 12, July, pp. 10121.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bernanke, B., 2010, “Central Bank Independence, Transparency, and Accountability” Speech at the Institute for Monetary and Economic Studies International Conference, Bank of Japan. Available at

    • Search Google Scholar
    • Export Citation
  • Bevilaqua, A., Mesquita, M. and A. Minella, 2007, “Brazil: Taming Inflation ExpectationsBCB Working Paper 129.

  • Bonomo, M. and B. Martins, 2016, “The Impact of Government-Driven Loans in the Monetary Transmission Mechanism: What Can We Learn From Firm-Level Data?BCB Working Paper No. 419.

    • Search Google Scholar
    • Export Citation
  • Calvo, G., 1978, “On the Time Consistency of Optimal Policy in Monetary Economy,” Econometrica, Vol. 46, November, pp. 141128.

  • Clinton, K., C. Freedman, M. Juillard, O. Kamenik, D. Laxton, and H. Wang, 2015, “Inflation-Forecast Targeting: Applying the Principle of Transparency,” June, IMF Working Paper 132.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Correa, A., M. Petrassi, R. Santos, 2016, “Price-Setting Behavior in Brazil: Survey EvidenceBCB Working Paper 422.

  • Cukierman, A., G. Miller and B. Neyapti, 2011, “Central Bank Reform, Liberalization and Inflation in Transition Economies—An International Perspective,” Journal of Monetary Economics.

    • Search Google Scholar
    • Export Citation
  • Dincer, N. and B. Eichengreen, 2014, “Central Bank Transparency and Independence: Updates and New Measures,” in International Journal of Central Banking, Vol. 38, No. 3, pp. 189253.

    • Search Google Scholar
    • Export Citation
  • Domit, S., C. Jackson and M. Roberts-Sklar, 2015, “Do Inflation Expectations Currently Pose a Risk to Inflation?Bank of England Quarterly Bulletin, Vol. 55, No. 2, pages 165180.

    • Search Google Scholar
    • Export Citation
  • Fischer, S., 2015, “Central Bank Independence” Herbert Stein Memorial Lecture National Economists Club, available at

    • Search Google Scholar
    • Export Citation
  • Freedman, C., and D. Laxton, 2009a, “Why Inflation Targeting?IMF Working Paper No. 09/86, Washington, DC: International Monetary Fund.

  • Freedman, C. and D. Laxton, 2009b, “Inflation Targeting Pillars: Transparency and Accountability,” IMF Working Paper No. 09/262, Washington, DC: International Monetary Fund.

    • Search Google Scholar
    • Export Citation
  • Friedman, B. and M. Woodford, 2011, “Handbook of Monetary Economics,” Vol. 3B, Chapter 22.

  • Goretti, M., and D. Laxton, 2005, “Long-Term Inflation Expectations and Credibility,” Box 4.2 in Chap. 4 of the September 2005 World Economic Outlook, International Monetary Fund.

    • Search Google Scholar
    • Export Citation
  • Gurkaynak, R., A. Levin, A. Marder, and E. Swanson, 2010, “Inflation Targeting and the Anchoring of Inflation Expectations in the Western Hemisphere,” Federal Reserve Bank of San Francisco Economic Review, pp. 2548.

    • Search Google Scholar
    • Export Citation
  • Hammond, G., 2012, “State of the Art of Inflation Targeting”, Bank of England Centre for Central Banking Studies.

  • Joyce, M., P. Lildholdt and S. Sorensen, 2010, “Extracting Inflation Expectations and Inflation Risk Premia from the Term Structure: A Joint Model of the UK Nominal and Real Yield Curves,” Journal of Banking & Finance, Vol. 34 February, pp. 28194.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kiley, M., 2007, “Is Moderate-to-High Inflation Inherently Unstable?International Journal of Central Banking.

  • Kumhof, M. and D. Laxton, 2007, “Improved Macroeconomic Performance—Good Luck or Good PoliciesBox 5.2 of Chap. 5, October 2007, World Economic Outlook, International Monetary Fund.

    • Search Google Scholar
    • Export Citation
  • Kydland, F. and E. Prescott, 1977, “Rules Rather than Discretion: The Inconsistency of Optimal Plans,” Journal of Political Economy, Vol. 85, June, pp. 47392.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Roache, S., 2014, “Inflation Persistence in Brazil—A Cross Country Comparison,” IMF Working Paper No. 14/55, Washington, DC, International Monetary Fund.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rosenberg, I., 2007, “Riksbank to Introduce Own Path for the Repo Rate," Speech at Danske Bank, Stockholm, available at

    • Search Google Scholar
    • Export Citation
  • Svensson, L., 2003, “Monetary Policy and Real Stabilization,” NBER Working Paper No. 9486.

  • Svensson and Woodford, 2003, “Implementing Optimal Policy through Inflation-Forecast Targeting” in Bernanke, B. and M. Woodford, eds., 2005, The Inflation-Targeting Debate, University of Chicago Press, pp. 983.

    • Search Google Scholar
    • Export Citation
  • Woodford, M., 2005, “Central-Bank Communication and Policy Effectiveness,” Proceedings—Economic Policy Symposium—Jackson Hole, Federal Reserve Bank of Kansas City, August.

    • Search Google Scholar
    • Export Citation
  • Yellen, J., 2012, “Revolution and Evolution in Central Bank Communications,” Speech at the Haas School of Business, University of California, Berkeley, Berkeley, California. Available at

    • Search Google Scholar
    • Export Citation

Prepared by Sílvia Domit, Douglas Laxton and Joannes Mongardini (RES).


Importantly, the new monetary policy regime was also supported by a strengthening of fiscal policy. For a comprehensive analysis of the early years of Brazil’s IT framework, please see Bevilaqua, Mesquita and Minella (2007).


See, for example, Svensson (2003).


See, for example, Kiley (2007) and Alves (2014).


Please see Freedman and Laxton (2009) for a discussion of the merits of IT and its intellectual roots.


The same conclusion holds if 2015 (year of great macroeconomic volatility in Brazil) is excluded from the sample.


As discussed in the IMF staff report for the 2014 Article IV Consultation and the 2013 Financial System Stability Assessment. Please see Bonomo and Martins (2016) for firm-level evidence on the impact of government-driven loans on the monetary transmission mechanism in Brazil.


The current formula for minimum wage, by affecting the growth in pension and other benefits, is also a source of fiscal pressure.


Whereas the central bank should look through the direct effects of such shocks, it would need to offset any second round effects, ie: ensure that the rise in the prices of a particular good does not lead to faster increases in the price of other goods and services.


Monetary policy has no impact on real variables such as output in the long run. The short-run trade-off has been extensively discussed in the monetary policy literature and is often represented by the Taylor efficiency frontier, which shows combinations of lowest volatility of inflation and output volatility for a given degree of interest rate volatility.


See Freedman and Laxton (2009), Kumhof and Laxton (2007) and Svensson (2003) for a discussion of how a credible monetary policy regime can result in improvements in inflation and output volatility.


See, for example, Svensson (2003) or Woodford (2005).


See, for example, Yellen (2012) and Woodford (2005).


Transparency is equally important to ensure the central bank remains accountable to the public for its actions.


See, for example, Woodford (2005) and Bernanke (2010).


While the policymaker does its best to minimize deviations from the point target, over long periods, inflation will be about equally likely to fall on either side of the target.


Some IFT central banks also add a band around their long-term point target as a way of explicitly communicating that inflation will deviate from the point target at times even if, on average, inflation will remain close to the point target. In contrast to the range target, the upper and lower bounds of the band are not the target.


Other reasons why forecast based on exogenous interest rates and exchange rates are inferior include concerns about appropriately measuring market expectations and technical difficulties regarding model solvency and stability. Please see Woodford (2005) and Rosenberg (2007) for comprehensive theoretical and practical explanations on why using exogenous paths is an inferior approach. Woodford (2005), in particular, discusses the experience of the Bank of England with exogenous interest rate assumptions.


Data availability is rich when it comes to inflation expectations of professional forecasters, but coverage of other sectors is limited. The most comprehensive sources are the BCB’s (Relatório Focus) and Consensus Economics (Long Term Consensus Forecasts) surveys, which cover professional forecasters’ expectations 1 to 5 years ahead. Market-based measures (derived from the difference between yields on nominal and inflation-linked government bonds) exist, but are not readily available to the general public (mainly due to the technical expertise required to compute estimates of inflation and liquidity risk premia). Households’ expectations are available from the FGV IBRE (Sondagem de Expectativas do Consumidor), but data only start in mid 2014. We are unaware of reliable measures of firms’ expectations.


Consensus Forecasts used to allow for international comparison, but BCB survey shows similar results. Longer-term inflation expectations measured at the 3-year-ahead horizon to ensure sample size is representative, but 5- year-ahead expectations exhibit similar pattern.


Dincer and Eichengreen (2014) construct an index of central bank transparency which acknowledges various dimensions of transparency covering five broad aspects: political, economic, procedural, policy and operation transparency. For each category, the authors establish three criteria, and award central banks a score of 0, ½ or 1 depending on the extent of transparency. Under this scoring system, the maximum score awarded to any central bank is 15 and the minimum is 0. The underlying data are gathered from central banks’ statutes, annual reports and other published documents. Whereas any measure of this nature is bound to be imperfect, the broad scoring patterns may be a reasonable approximation of international trends.


By central bank independence, we refer specifically to independence from political influence in setting policy. The Banco Central do Brasil has autonomy to make monetary policy decisions.


Decree no. 3088 (21 June, 1999) only goes as far as establishing the inflation-targeting framework. For examples of central bank acts, please see:, or


Please see Decree no. 3088 (21 June, 1999).


For a discussion of how central banks have developed an FPAS to support IFT regimes, please see Clinton and others (2015).


Another available tool is publishing endogenous interest rate paths (section B).


For example, a discussion of the evolution of labor supply, productivity, cost pressures, wages, margins, savings, the degree of slack in the economy, pass-through from price level shocks, foreign demand and prices, etc.


The benefits from improving the frontier itself are not captured in this exercise.


The authors would like to thank Sarma Jayanthi for his assistance with this section.

Brazil: Selected Issues
Author: International Monetary Fund. Western Hemisphere Dept.