Articulo 5

IMF Country Report No.
17/216
BRAZIL
SELECTED ISSUES
July 2017
This paper on Brazil was prepared by a staff team of the International Monetary Fund as
background documentation for the periodic consultation with the member country. It is
based on the information available at the time it was completed on June 20, 2017.
Copies of this report are available to the public from
International Monetary Fund • Publication Services

PO Box 92780 • Washington, D.C. 20090
Telephone: (202) 623-7430 • Fax: (202) 623-7201
E-mail: publications@imf.org Web: http://www.imf.org
Price: $18.00 per printed copy
International Monetary Fund

Washington, D.C.
© 2017 International Monetary Fund

BRAZIL
SELECTED ISSUES
June 20, 2017
Approved By Prepared by H. Barbosa, C. Góes, N. Biljanovska, I. Karpowicz,

Western Hemisphere T. Matheson (all WHD), I. Krznar (MCM), C. Mulas-Granados
Department (FAD), V. Malta, M. Mendes Tavares, and K. Moriyama (all
SPR).
CONTENTS
INVESTMENT IN BRAZIL: FROM CRISIS TO RECOVERY______________________________ 5

A. Introduction _________________________________________________________________________ 5
B. Aggregate Empirical Analysis_________________________________________________________ 7
C. A Closer Look at Investment at the Firm Level _______________________________________ 9
D. Summary and Conclusion ___________________________________________________________ 12
FIGURES
1. Gross Fixed Capital Formation and Other Key Variables _____________________________ 6
2. Decomposition of Gross Fixed Capital Formation Growth ____________________________ 9
TABLES
1. Variable Definitions__________________________________________________________________ 11
2. Regression Results, All Sectors ______________________________________________________ 12
3. Regression Results, by Sector _______________________________________________________ 13
APPENDIX
I. Data, Results, and Robustness _______________________________________________________ 14
References _____________________________________________________________________________ 18
BRAZIL’S BUSINESS ENVIRONMENT AND EXTERNAL COMPETITIVENESS _______ 19

A. Ease of Doing Business ______________________________________________________________ 19
B. Recent Reforms Affecting the DB Score _____________________________________________ 23
C. Announced Micro Reform Measures ________________________________________________ 24
BRAZIL
D. External Competitiveness ___________________________________________________________ 26

E. Factors Explaining Deteriorating Competitiveness __________________________________ 27
F. Benefits of Furthering Trade Liberalization in Brazil _________________________________ 32
G. Conclusions _________________________________________________________________________ 34
BOXES
1. History of Trade Liberalization in Brazil______________________________________________ 28
2. Literature on the Impact of Trade Liberalization on Productivity and Growth _______ 32
FIGURES
1. Investment and Tariffs: Brazil and Emerging Markets _______________________________ 29
2. Global Competitiveness: Brazil and Emerging Markets ______________________________ 31
References _____________________________________________________________________________ 35
DISTRIBUTIONAL EFFECTS OF BRAZIL’S PENSION REFORM _______________________ 37

A. The Reform of Social Security _______________________________________________________ 37
B. Stylized Facts on the Social Security System and Inequality _________________________ 38
C. Some Distributive Effects of Pension Reform _______________________________________ 41
APPENDIX
I Figure and Table ______________________________________________________________________ 45
References _____________________________________________________________________________ 48
INEQUALITY IN BRAZIL: A MICRO-DATA ANALYSIS _______________________________ 49

A. Overview ____________________________________________________________________________ 49
B. Historical Trends in Regional Inequality 2004–14 ___________________________________ 50
C. Macro-Policies and Inequality Outcomes: Regression Analysis _____________________ 56
D. Conclusions _________________________________________________________________________ 63
BOXES
1. The Cost of Living Adjustment ______________________________________________________ 51
2. Returns to Education and Public-Private Wage Gap_________________________________ 58
3. An Example of Poorly Targeted Transfers: Public Universities _______________________ 60
FIGURES
1. Convergence in the Consumption of Goods by Households ________________________ 51
2. Income Inequality in Brazilian States: A Dynamic Decade ___________________________ 53
2 INTERNATIONAL MONETARY FUND

BRAZIL
3. Income Inequality in Brazilian States: Some Evidence of Convergence ______________ 54

4. Income Inequality in Brazilian States – Disaggregated ______________________________ 55
5. Income Inequality in Brazilian States – Convergence in the Middle _________________ 56
TABLE
1. Coefficients from the 2SLS Procedure _______________________________________________ 62
APPENDIX
I. Data and Sources ____________________________________________________________________ 64
References _____________________________________________________________________________ 71
INTEREST RATES AND INFLATION IN BRAZIL ______________________________________ 74

A. Introduction _________________________________________________________________________ 74
B. Empirical Analysis ___________________________________________________________________ 75
C. How can Lower Inflation Be Achieved over the Long Term? ________________________ 77
D. Summary and Discussion ___________________________________________________________ 78
FIGURES
1. Headline Inflation and Policy Rate ___________________________________________________ 75
2. Correlation: Headline Inflation and Interest Rate ____________________________________ 76
3. Headline Inflation after 100bps Cut in Policy Rate __________________________________ 77
4. Simulated Responses to a Change in the Inflation Target from 4.5 Percent to
2 Percent _______________________________________________________________________________ 78
5. Disinflation Episodes Across Countries ______________________________________________ 79
APPENDIX
I. Data and Robustness ________________________________________________________________ 81
References _____________________________________________________________________________ 85
THE EFFECT OF FISCAL CONSOLIDATION ON REAL INTEREST RATES IN

EMERGING MARKET ECONOMIES ___________________________________________________ 86
A. Fiscal Policies in Brazil _______________________________________________________________ 86
B. Fiscal Consolidation Episodes _______________________________________________________ 87
C. Econometric Specification ___________________________________________________________ 88
D. The Effect of Debt Outstanding _____________________________________________________ 91
E. Robustness __________________________________________________________________________ 92
F. Conclusion and Policy Recommendations ___________________________________________ 93
INTERNATIONAL MONETARY FUND 3

BRAZIL
FIGURES
1. Cyclically Adjusted Primary Balance _________________________________________________ 86
2. Real Rates in a Set of EMs ___________________________________________________________ 87
3. The Response of Real Interest Rates to Fiscal Consolidation Episodes in EMs ______ 90
4. Fiscal Adjustment Reversals _________________________________________________________ 90
5. The Response of Real Interest Rates to Fiscal Consolidation Episodes in EMs (D>60
percent of GDP) ________________________________________________________________________ 91
6. The Response of Real Interest Rates to Fiscal Consolidation Episodes in EMs
(Restricted sample to EMs only) _______________________________________________________ 93
APPENDICES
I. Data __________________________________________________________________________________ 94
II. Methodology ________________________________________________________________________ 97
References _____________________________________________________________________________ 98
BOOSTING PRODUCTIVITY: TAXES AND RESOURCE MISALLOCATION IN

BRAZIL ________________________________________________________________________________ 99
A. The Productivity Challenge in Brazil _________________________________________________ 99
B. Resource Misallocation in Brazil ____________________________________________________ 100
C. Upgrading the Tax System to Reduce Resource Misallocation _____________________ 103
D. Conclusions and Policy Recommendations ________________________________________ 108
FIGURES
1. Labor Productivity Growth in Selected Countries____________________________________ 99
2. Distribution of Firm-Level Revenue Productivities __________________________________ 101
3. Cross-Country Resource Allocation Efficiency ______________________________________ 102
4. Growth Impact from Improving Resource Allocation Efficiency ____________________ 103
5. GDP Growth Effects from Reducing Tax Disparity __________________________________ 105
6. GDP Growth Effects from Reducing Corporate Debt Bias __________________________ 107
7. GDP Growth Effects from Reducing Tax Informality ________________________________ 108
References ____________________________________________________________________________ 110

BRAZIL
INVESTMENT IN BRAZIL: FROM CRISIS TO

RECOVERY1
While Brazil’s deep recession has been broad based, it has been marked by a particularly
large fall in investment. Real investment fell by around 30 percent between the beginning
of 2014 and the beginning of 2017. This chapter finds that a variety of factors contributed
to the investment decline, including a deterioration in Brazil’s medium-term growth
prospects, rising real interest rates, falling terms of trade, rising uncertainty related to
economic policy, rising levels of corporate leverage and lower cash flow. Some of the
factors that have weighed on investment over recent years have begun to normalize
providing some impetus for a recovery. However, still-high levels of corporate leverage
and the prospect of continued uncertainty related to economic policy settings suggest a
turnaround in investment is likely to be subdued.
A. Introduction
1. Brazil’s deep recession has been broad based with Brazil: National Accounts Components
(Index, 2010Q1 = 100)
a particularly large fall in investment. Real investment 125
120
recovered strongly following the global financial crisis amid 115
rising terms of trade, low interest rates, rapidly expanding 110
consumption, and widespread optimism about Brazilian 105

100
growth prospects. However, the expansion proved short 95 Real GDP
lived as the favorable factors that supported the recovery 90 Private consumption
Government consumption
began to wane and economic and policy uncertainty began 85
Gross fixed capital formation
80
to rise. Since the beginning of 2014, real investment has
Sep-10
Sep-11
Sep-12
Sep-13
Sep-14
Sep-15
Sep-16
Mar-10
Mar-11
Mar-12
Mar-13
Mar-14
Mar-15
Mar-16
contracted by around 30 percent, far weaker investment than
most other emerging markets experienced over this time. Source: Haver Analytics.
2. Some of the factors that are hampering

Brazil and EMs: Real Gross Fixed Capital Formation
investment might soon support a recovery, but other (Index, 2000 = 100)
370
factors—such as high levels of corporate leverage and EM Interquartile range
320 EM Median
policy uncertainty—suggest the recovery might be more Brazil
prolonged (Figure 1). The central bank has begun an easing 270
cycle and real interest rates are expected to fall significantly 220
over the coming year, and, after trending down for several 170
years, Brazil’s terms of trade stabilized and improved in early 120
2016. By reducing funding costs and increasing profitability, 70

both factors should help to support a recovery in investment.
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
However, these developments come against the backdrop of Source: World Economic Outlook.
a prolonged recession that has damaged Brazil’s long-term
1
Prepared by Ivo Krznar (MCM) and Troy Matheson (WHD).

BRAZIL
growth prospects and led to higher levels of corporate leverage; prospects of weaker growth in the
future hurts expected returns and discourages investment while higher leverage reduces the
demand and supply of investment funding. Moreover, following a dramatic deterioration in the fiscal
position and the impeachment of President Rousseff, the government has embarked on an
ambitious fiscal reform agenda that is subject to implementation challenges, contributing to
heightened policy uncertainty that could weigh on firms’ willingness invest.
Figure 1. Brazil: Gross Fixed Capital Formation and Other Key Variables
Gross Fixed Capital Formation

(Index, 2000=100; and percent of GDP, respectively)
200 22
180 21
20
160
19
140
18
120
17
100 16
Level GFKF GFKF / GDP, rhs
80 15
2000Q1
2001Q1
2002Q1
2003Q1
2004Q1
2005Q1
2006Q1
2007Q1
2008Q1
2009Q1
2010Q1
2011Q1
2012Q1
2013Q1
2014Q1
2015Q1
2016Q1
Real Interest rate and GFCF Terms of Trade Index and GFCF
18 22 135 22
16 Terms of Trade
21 130 21
GFKF / GDP, rhs
14 125
20 20
12 120
10 19 115 19
8 18 110 18
6 105
17 17
4 100
2 16 16
95
Real Interest Rate GFKF / GDP, rhs
0 15 90 15
2000Q1
2001Q1
2002Q1
2003Q1
2004Q1
2005Q1
2006Q1
2007Q1
2008Q1
2009Q1
2010Q1
2011Q1
2012Q1
2013Q1
2014Q1
2015Q1
2016Q1
2000Q1
2001Q1
2002Q1
2003Q1
2004Q1
2005Q1
2006Q1
2007Q1
2008Q1
2009Q1
2010Q1
2011Q1
2012Q1
2013Q1
2014Q1
2015Q1
2016Q1
Economic Policy Uncertainty Index and GFCF Corporate Leverage (debt to GDP) and GFCF
350 22 50 Corporate Leverage 22
Policy Uncertainty
300 21 45 21
GFKF / GDP, rhs GFKF / GDP, rhs
20 20
250 40
19 19
200 35
18 18
150 30
17 17
100 16 25 16
50 15 20 15
2000Q1
2001Q1
2002Q1
2003Q1
2004Q1
2005Q1
2006Q1
2007Q1
2008Q1
2009Q1
2010Q1
2011Q1
2012Q1
2013Q1
2014Q1
2015Q1
2016Q1
2000Q1
2001Q1
2002Q1
2003Q1
2004Q1
2005Q1
2006Q1
2007Q1
2008Q1
2009Q1
2010Q1
2011Q1
2012Q1
2013Q1
2014Q1
2015Q1
2016Q1
Source: Staff estimates and Haver Analytics.

BRAZIL
3. This chapter examines the proximate causes of dramatic fall in investment in Brazil
and the prospects for investment going forward. Section B provides an analysis and discussion of
Brazil’s investment using aggregate data. Firm-level data are examined and discussed in Section C.
The chapter concludes with a discussion of the prospects for a recovery in investment.
B. Aggregate Empirical Analysis
4. Medium-term growth and investment are linked. Current levels of investment increase
the economy’s capital stock and enhance future labor productivity and growth. At the same time,
firms’ expectations of medium-term growth help to determine expected returns to investment and
investment levels themselves. As such, the empirical specification is chosen to allow for feedback
effects from growth expectations and investment and for the impact of other variables that influence
investment decisions, including input costs (regulated prices and unit labor costs), funding costs (the
real interest rate), profitability (terms of trade), leverage, and economic policy uncertainty.2
5. Two equations are jointly estimated to describe the behavior of quarterly real
investment growth. The first equation determines the part of medium-term growth expectations
(3-years-ahead) that cannot be explained by current economic conditions and the second equation
determines real investment growth. Medium-term growth expectations (3-years-ahead) are
assumed to be determined by annual real GDP ( growth, annual regulated-price ( inflation,
annual changes in the
Data, Sources and Transforms
terms of trade
Label Variable Source Units
and the real Real GDP Growth Expectations (3-years ahead) BCB Level, %
Real Gross Fixed Capital Formation IBGE Level, log
interest rate ( , and Real GDP IBGE Level, log
annual growth in Regulated Prices, IPCA IBGE Level, log
Term of Trade Haver Level, log
equity prices ( and Real Interest Rate (Selic - 12-month-ahead inflation expectations) BCB Level, log
unit labor costs ( Equity Price Index Haver Level, log
Unit Labor Costs, $R BCB Level, log
(see table for data Corporate Leverage (corporate debt to GDP) IMF Level, %GDP
sources and Economic Policy Uncertainty Haver Level, log
definitions):3
∆ ∆ ∆ ∆ ∆ ∆ (1)
where ∆ represents the annual change in each variable and is the part of medium-term growth
expectations that represents other factors unrelated to current economic conditions. These factors
represent an autonomous reassessment of medium-term growth, above and beyond the impact of
contemporaneous (short term) economic developments. Quarterly investment growth is assumed to
be driven by the same variables as growth expectations, in addition to the estimated residual from
2
Economic policy uncertainty is measured using the Economic Policy Uncertainty (EPU) Index for Brazil in the same
manner as the newspaper-based EPU Index for the United States, following the methods in "Measuring Economic
Policy Uncertainty" by Baker, Bloom and Davis (see policyuncertainty.com and Appendix I, Section A for more details).
3
The estimates of medium-term growth expectations are described in Appendix I, Section B.

BRAZIL
equation (1) and lags of corporate leverage (measured as total corporate debt over GDP) and an
index of economic policy uncertainty :4
∆
∆ ∆ ∆ ∆ ∆ ∆ ∑ ∑ 2
where ∆ represents the quarterly change in each variable.
6. The estimated model provides a good description of behavior of investment over

history. Three different variants of the model are estimated using Seemingly Unrelated Regressions
(the estimation results are displayed in Appendix I, Section B). The preferred model specification
(model 3) fits quarterly investment growth very well considering the volatility of the series (the
adjusted R-squared statistic is around 0.7). Moreover, the signs of the coefficients in the investment
equation are intuitive. Specifically, the estimated parameters suggest:
 Investment increases with higher autonomous growth expectations for the future and higher
terms of trade;
 Investment decreases with higher real interest rates, unit labor costs, regulated prices, leverage,
and policy uncertainty.
7. What explains the rapid drop in investment since 2014? Estimates suggest that
developments hampering investment over this period include a rise in costs (chiefly a sharp increase
in regulated prices, such as energy prices, but also unit labor costs), falling terms of trade—
impacting prospects for commodity exporters—and higher interest rates (Figure 2). The
“autonomous” part of the deterioration in the medium-term outlook for growth and heightened
policy uncertainty have been the most significant drags on investment over this time, with each
factor reducing investment by around 10 percent since beginning of 2014. Higher leverage
contributed to fall in investment primarily during the period when it was rising fast (2015).
4
Note, variants of the model that include leverage and policy uncertainty in equation 1 were also estimated but
leverage and policy uncertainty were not found to be statistically significant.

BRAZIL
Figure 2. Brazil: Decomposition of Gross Fixed Capital Formation Growth

(Quarter-on-quarter growth accumulated since 2014Q1, deviations from avg, all variables demeaned)
10
-10
-20
-30
Regulated Prices Terms of Trade
Real interest rate Unit labor cost
-40 Autonomous mid-term growth exp. Leverage
Uncertainty Other factors
-50
2014Q1
2014Q2
2014Q3
2014Q4
2015Q1
2015Q2
2015Q3
2015Q4
2016Q1
2016Q2
2016Q3
2016Q4
Source: IMF staff estimates.
C. A Closer Look at Investment at the Firm Level
8. As in Magud and Sosa (2015) and Li and others (2015), a panel regression model that
relates each firm’s investment-to-capital ratio to several determinants of firms’ investment
decisions was estimated to study factors driving investment at the individual firm level. The
macroeconomic determinants of investment studied in the previous section are assumed to affect
investment via firms’ balance sheets. The model differentiates between fundamental factors that
drive investment decisions (factors that capture the marginal productivity of capital) and financial
factors that can also affect investment (factors that capture financing constraints). The investment-
to-capital ratio is assumed to be driven by:
Fundamentals:
 The change in sales to proxy for demand and expected future growth (as in the standard
accelerator model of investment);
 Tobin’s Q to capture the expected marginal return on investment.
Financial Factors:
 Leverage, cash flow, the change in debt, and debt repayment capacity.

BRAZIL
The only other control variable included is a measure of political uncertainty; this variable is less
likely to be endogenous to the firm-level variables than other macroeconomic variables such as
interest rates, GDP, and the terms of trade.5
9. Leverage, cash flow, and changes in debt capture the relevance of each firm’s financial
structure for access to investment finance. High leverage may constrain firms’ ability to obtain
external financing for new investment; high leverage provides less investment incentives to
controlling shareholders as a larger share of the gains will necessarily accrue to debtholders. Cash
flow is commonly used in investment models as an indicator for internally available funds (see
Hubbard, 1998) and the severity of financial constraints—tighter financial constraints increase
reliance on internal funding for investment. While higher leverage is expected to be negatively
associated with investment, the flow of debt is expected to be positively related to capital
expenditures because financing investment is one of the main reasons to incur new debt. Tighter
debt repayment capacity, on the other hand, is expected to be associated with lower investment.
10. The baseline model specification is:
where is firm i’s net investment in period t, its capital stock at the beginning of the period,
LEV is the firm’s leverage, Δ is its change in total debt, Δ is the change total sales,
_ its average Tobin’s Q (a proxy for unobservable marginal Q), is a measure of the
firm’s interest coverage ratio (a proxy for debt repayment capacity), and is the index of
political uncertainty used in the previous section. To control for unobserved heterogeneity across
6,7
firms, and other aggregate effects not explicitly modeled here, the model is estimated with firm-
specific ( ) and time effects ( ); firm-specific effects control for systematic differences in the
average investment rate across firms and time effects control for a common investment component
reflecting other macroeconomic factors that can influence firm-level investment. The final term in
the equation is the idiosyncratic error, .8
5
For example, expected GDP growth rate will be reflected in Tobin’s Q, terms of trade will affect cash flow, Tobin’s Q,
sales; interest rates will affect cash flow and the interest coverage ratio. Specifications were examined that included a
real interest rate and a nominal interest rate, but the associated coefficients were insignificant.
6
The empirical investment literature shows that lagged investment rate might be an important determinant of
current investment spending (Gilchrist and Himmelberg, 1995; Eberly and others, 2012). In the case of Brazil,
including lagged investment ratio as an explanatory variable did not significantly change the estimation results.
7
Different interaction terms were also explored such as between leverage and firm size ( smaller firms tend to be
more dependent on bank financing and have lower spare capacities and a lower ability to access alternative financing
options leverage and uncertainty(firms with relatively higher leverage reduce investment more aggressively in
response to higher uncertainty shock), cash flow and uncertainty (higher uncertainty could increase or decrease the
marginal propensity to invest out of cash flows). None of these interaction terms were significant probably due to
multicollinearity issues.
8
Leverage enters the equation with a lag whereas, cash flow, debt growth and the interest coverage ratio
contemporaneously, as in Magud and Sosa (2015) and Li and others (2015).

BRAZIL
11. Data from firm-level balance sheets and income statements are sourced from Capital
IQ. The sample covers over 4,000 Brazilian firms spanning 1995 to 2016.9 To adjust for outliers, all
firm-specific ratios are winsorized using 1st and 99th percentiles and some additional constraints are
imposed.10 The definition of each variable is provided in Table 1.
Table 1. Brazil: Variable Definitions

Variable Definition Robustness check
Capital expenditure within the year over the capital stock at the beginning
Investment ratio of the year (net property, plant and equipment+depreciation and
amortization-investment within the year)
Total debt net of cash and
Total debt net of cash+cash equivalents over total assets-cash and cash
Leverage ratio cash equivalents over total
equivalents
common equity or EBITDA
Net income+depreciation over the capital stock at the beginning of the EBIT over the capital stock at
Cash flow ratio
year the beginning of the year
Change in debt ratio Change in total debt over over the capital stock at the beginning of the year
Change in sales ratio Change in revenues over the capital stock at the beginning of the year
Tobin's Q Market capitalization-total equity+total assets over total assets Price to book ratio
Interest coverage ratio EBITDA over interest expense
Change in Economic Policy Uncertainty by Scott Baker, Nicholas Bloom

Uncertainty
and Steven J. Davis
Source: IMF staff estimates.
12. Estimation results of the baseline model are consistent with the findings of the
analysis at the aggregate level. Table 2 shows the estimated parameters (and their statistical
significance) for 7 variations of the model, ranging from a simple model that relates investment to
leverage only to the full model specification described above. Robustness checks, including models
using the alternative variable definitions displayed in Table 2, are displayed in Appendix I, Section C.
The results can be summarized as follows:
 Investment increases with higher expected future profitability (proxied by Tobin’s Q and sales
growth), cash flows, and debt flows, as expected;
 Investment decreases with higher leverage and higher political uncertainty. For example, a
10 percentage point rise in a large firm’s leverage is associated with a 1.1‒1.4 percentage point
fall in the investment-to-capital ratio. This suggests that financial constraints and political
uncertainty play an important role in investment decisions and could help to explain the recent
large drop in investment at the aggregate level.
9
While this is potentially a large sample, the data is sparse and concentrated in the period from 2010 to 2016.
10
The cutoff values are 250 and -50 for investment to capital ratios, 20 and -20 for sales growth ratios and debt
growth ratios, 10 and -10 for cash flow ratios, 50 and 0 for Tobin’s Q and 100 and 0 for leverage ratios.

BRAZIL
Table 2. Brazil: Regression Results, All Sectors

Dependent variable: Investment to capital ratio
Explanatory variables (1) (2) (3) (4) (5) (6) (7)
Leverage, lag -0.116 *** -0.120 *** -0.115 *** -0.110 *** -0.139 *** -0.116 *** -0.112 ***
(0.022) (0.023) (0.028) (0.021) (0.019) (0.022) (0.021)
Cash Flow ratio 0.031 *** 0.029 *** 0.022 *** 0.034 ** 0.049 *** 0.049 ***
(0.005) (0.005) (0.005) (0.015) (0.015) (0.015)
Change in debt ratio 0.016 *** 0.011 *** 0.007 0.008 0.008 *
(0.002) (0.003) (0.005) (0.005) (0.005)
Change in sales ratio 0.010 *** 0.006 0.003 0.003
(0.003) (0.006) (0.007) (0.007)
Tobin's Q 0.005 *** 0.005 *** 0.004 ***
(0.002) (0.002) (0.002)
Interest coverage ratio 0.001 0.001
(0.001) (0.001)
Uncertainty, diff, lag -0.001 **
(0.000)
Number of observations 10,763 10,378 10,144 9,805 2,392 2,311 2303

Number of firms 3,286 3,151 3,052 2,930 297 291 291
2
R 0.01 0.03 0.03 0.03 0.04 0.05 0.05
Source: Author's calculations
Note: ***, **, * indicate significance at the 1 percent, 5 percent, and 10 percent levels, respectively.
Standard errors in parentheses are Driscoll and Kraay (1998) robust to heteroscedasticity, autorcorrelation with MA(q),
and cross-sectional dependency. The estimation period is 1995–2016. Firm-level fixed effects and time effects are included.
13. Higher leverage and policy uncertainty are negatively correlated with investment
across most sectors of the economy. Table 3 displays results from models estimated using firms
from eight different sectors of the economy. Leverage is statistically significant across most sectors,
with health and real estate sectors being the only exceptions. The impact of leverage on investment
is particularly large in the energy sector, possibly reflecting the relatively high levels of leverage and
falling investment levels at the state-owned energy producer, Petrobras. At the same time, policy
uncertainty appears to have significant effects on investment levels in the healthcare, industrials, real
estate, and utility sectors, possibly due to a higher aversion to risks associated macroeconomic
policy settings among these industries.
D. Summary and Conclusion

14. There are several factors that have contributed to the decline in investment since
2014. Rising costs, falling profitability—and expected profitability—and higher levels of corporate
leverage appear to have played a role in the decline in investment over the course of the recession.
The empirical evidence in this chapter also suggests that a rise in policy uncertainty has played a
significant role in the decline.

BRAZIL
Table 3. Brazil: Regression Results, by Sector

Consumer Consumer Real Telecomm-

Energy Healthcare Industrials Utilities
Discretionary Staples Estate unication
Explanatory variable
Leverage, lag -0.142 *** -0.230 *** -0.336 *** 0.072 -0.119 *** -0.062 -0.369 *** -0.182 ***
(0.047) (0.033) (0.069) (0.120) (0.036) (0.042) (0.135) (0.050)
Cash Flow ratio -0.016 0.069 *** 0.094 *** -0.056 0.029 * 0.039 * 0.099 *** -0.001
(0.017) (0.020) (0.025) (0.043) (0.018) (0.022) (0.035) (0.010)
Change in debt ratio 0.002 -0.001 0.044 ** -0.002 0.019 *** 0.001 0.101 *** -0.003
(0.004) (0.006) (0.022) (0.015) (0.006) (0.005) (0.038) (0.004)
Change in sales ratio 0.017 *** 0.007 ** -0.019 ** -0.005 0.007 ** 0.008 -0.025 0.016 ***
(0.007) (0.003) (0.007) (0.014) (0.007) (0.005) (0.019) (0.003)
Uncertainty, diff, lag -0.000 -0.000 -0.000 -0.001 *** -0.000 *** -0.001 *** -0.001 -0.000 ***
(0.000) (0.000) (0.000) (0.008) (0.000) (0.000) (0.001) (0.000)
Number of observatio 1,082 970 282 329 2,160 1,133 137 1215
Number of firms 357 272 80 128 677 422 28 359
R2 0.05 0.08 0.21 0.06 0.06 0.04 0.29 0.06
15. Some factors point to stronger investment in the short term. The empirical results
suggest that stabilization of regulated-price inflation, the terms of trade, and real interest rates
should improve both investment growth and growth expectations. The large and necessary
increases in fuel and electricity tariffs in 2015 and are not expected to be repeated going forward,
the central bank has begun an easing cycle that is expected to continue over the remainder of 2017,
and Brazil’s terms of trade have improved markedly since the beginning of 2016. These are positive
signs.
16. However, headwinds remain. While stabilization of some of the factors weighing on
investment should provide some support investment in the short term, the prospect of a return to
past levels of strong investment growth crucially depends on an alleviation of other sources of
weakness, most notably high levels of corporate leverage and policy uncertainty and low medium-
term growth expectations. While corporate leverage appears to have stabilized recently, it remains
high from both a historical perspective and relative to other countries in the region. The analysis
also suggests that if medium-term growth expectations remain subdued, the recovery in investment
may be incomplete. Likewise, implementation challenges related the governments reform agenda—
most notably related to pension reform—and the prospect of rising levels of government debt over
the next several years could lead to persistently high policy uncertainty that might weigh on
investment decisions for some time.

BRAZIL
Appendix I. Data, Results, and Robustness
A. Economic Policy Uncertainty
The Economic Policy Uncertainty (EPU) Index for Brazil is computed in the same manner as the
newspaper-based EPU Index for the United States, following the methods in "Measuring Economic
Policy Uncertainty" by Baker, Bloom and Davis (policyuncertainty.com).
 The index is derived from the text archives of the newspaper Folha de São Paulo from 1991
onwards. The raw data for the index is the number articles containing relevant terms.
Specifically, the number of articles containing the terms "incerto" or "incerteza", "econômico" or
"economia", and one or more of the following policy-relevant terms: regulação, déficit,
orçamento, imposto, banco central, alvorada, planalto, congresso, senado, câmara dos
deputados, legislação, lei, tarifa.
 To obtain the EPU rate, the raw EPU counts are scaled by the number of all articles in the same
newspaper and month. The resulting series is multiplicatively rescaled to a mean of 100 from
January 1991 to December 2011.
B. Estimating Medium-Term Growth Expectations
The Brazilian central bank has a very comprehensive consensus expectations survey. On a
weekly basis, forecasters submit their expectation for end-of-year Real GDP growth for the current
year, as well as for the 4 years ahead. The Central Bank then reports averages, medians, and other
properties of the sample in their website.
While there is no constant forecast horizon, forecasts usually behave in a predictable fashion
(see Figure below). The surveyors ask analysts for end-of-year forecasts, which means that, for any
given year, the forecast horizon shortens as time passes. However, since the underlying models
expect the economy to return to potential GDP growth over the long run, there is a strong
relationship between expected growth and forecast horizons.
This exercise builds on this relationship. The starting point to estimate a continuous time-series
for mid-term growth expectations, then, regresses forecasts on a polynomial of the forecast horizon:
Δ , , 1

BRAZIL
where Δ is end-of-year growth forecast for date s at time t; , denotes the forecast
horizon in days, , and are time varying parameters estimated for each period t. Those
parameters yield a continuous time series for growth expectations for any horizon h ahead:
Δ ̂ 2
Brazil: Consensus GDP Growth Forecasts

(X-axis: forecast horizon, in years ahead; Y-axis: Yearly GDP
Growth median forecast)
6
5 R² = 0.98
4
R² = 0.98
3
R² = 0.99
R² = 0.98
2
1 R² = 0.99
0
-1 Dec-06 Dec-08 Dec-11
-2 Dec-13 Dec-15
-3
0 1 2 3 4 5
Source: Central Bank of Brazil.

BRAZIL
C. Estimation Results: Aggregate Model
Table 1. Brazil: Results—Aggregate Model

Model
(1) (2) (3)
Growth Expectations
∆4 0.069** 0.093*** 0.093***
(0.026) (0.027) (0.027)
∆4 -0.022 -0.031
(0.019) (0.020)
∆4 0.030** 0.028* 0.029**
(0.014) (0.014) (0.014)
∆4 -0.094*** -0.105*** -0.106***
(0.035) (0.033) (0.032)
∆4 -0.006 -0.009** -0.010***
(0.004) (0.004) (0.004)
∆4 0.013 0.011
(0.015) (0.015)
R-Squared 0.507 0.499 0.496
Adjusted R-Squared 0.458 0.439 0.467
Investment Growth
Δ 1 -0.793*** -0.512*** -0.542***
(0.187) (0.167) (0.165)
Δ 0.411*** 0.448*** 0.476***
(0.108) (0.100) (0.092)
Δ 1 -0.009*** -0.009*** -0.009***
(0.003) (0.002) (0.002)
Δ 1 0.096*** 0.041
(0.029) (0.028)
Δ 1 -0.104 -0.167** -0.173***
(0.068) (0.064) (0.059)
1 1.360** 1.110** 1.179**
(0.536) (0.479) (0.483)
1 0.520*** 0.529***
(0.136) (0.130)
2 -0.381* -0.581***
(0.206) (0.131)
3 -0.282
(0.212)
4 0.080
(0.143)
1 0.143
(0.986)
2 0.201
(1.047)
3 -1.976* -1.370*
(1.039) (0.821)
4 -2.067** -2.181**
(0.951) (0.877)
R-Squared 0.696 0.790 0.770
Adjusted R-Squared 0.635 0.698 0.704
Jointly estimated standard errors in parentheses. * p < 0.1, ** p <
0.05, *** p < 0.01

BRAZIL
D. Robustness Checks: Firm-Level Model
Table 2. Brazil: Results—Firm-Level Model, Robustness Checks

(4) with
(4) with leverage
leverage (net (7)
(net debt/EBITD (4) with (5) with estimated
Explanatory variables debt/equity) A) cash ratio Tobin Q using GMM
Leverage, lag -0.003 *** -0.001 ** -0.110 *** -0.292 *** -0.172 ***
(0.001) (0.000) (0.021) (0.046) (0.068)
Cash Flow ratio 0.028 *** 0.034 *** 0.021 ** 0.025 0.049 ***
(0.007) (0.005) (0.005) (0.016) (0.014)
Change in debt ratio 0.014 *** 0.015 *** 0.009 *** 0.009 ** 0.008
(0.003) (0.002) (0.002) (0.005) (0.007)
Change in sales ratio 0.008 *** 0.011 *** 0.009 *** 0.006 0.005
(0.003) (0.003) (0.002) (0.006) (0.008)
Tobin's Q 0.004 *** 0.005 ***
(0.001) (0.004)
Interest coverage ratio 0.001
(0.001)
Uncertainty, diff, lag 0.000 **
(0.000)
Number of observations 9,144 11,391 9,779 2,632 2,190

Number of firms 2,833 3,357 2,925 382 290
R2 0.03 0.04 0.03 0.06 -

BRAZIL
References
Driscoll, John and Aart Kraay, 1998, “Consistent Covariance Matrix Estimation with Spatially
Dependent Panel Data,” The Review of Economics of Economics and Statistics, Vol. 80(4), pp. 549−60.
Eberly, Janice, Sergio Rebelo, and Nicolas Vincent, 2012, “What Explains the Lagged-Investment
Effect?” Journal of Monetary Economics, Vol. 59, pp. 370−80.
Gilchrist, Simon and Charles Himmelberg, 1995, “Evidence on the Role of Cash Flow for Investment,”
Journal of Monetary Economics, Vol. 36, pp. 541−72.
Li, Delong, Nicolas Magud, and Fabian Valencia, 2015, “Corporate Investment in Emerging Markets:
Financing vs. Real Options Channels,” IMF Working Paper No. 15/285 (Washington: International
Monetary Fund).
Magud, Nicolas and Sebastian Sosa, 2015, “Investment in Emerging Markets: We are Not in Kansas
Anymore…Or Are We?” IMF Working Paper No. 15/77 (Washington: International Monetary Fund).

BRAZIL
BRAZIL’S BUSINESS ENVIRONMENT AND EXTERNAL

COMPETITIVENESS1
This chapter provides an overview of competitiveness challenges in Brazil and identifies
policies to improve competitiveness; it documents the impact of reforms on
competitiveness using the parameters from existing empirical and quantitative studies,
and reports on priority areas in need of reform flagged by the World Bank’s Ease of Doing
Business (DB) assessment. The analysis underscores the importance of reducing tariffs as
well as non-tariff barriers. Lower tariffs on capital goods, greater integration into global
value chains, more flexible labor, product and financial markets, and lighter procedures
to pay taxes and obtain business permits are important for improving the external
competitiveness.
A. Ease of Doing Business
1. Brazil’s position in the World Bank DB ranking is disappointing, but has improved
marginally since 2010. In 2017, Brazil ranked 123rd in a total of 190 states on which data were
collected. Brazil’s 2017 ranking is better compared to 2010, when it occupied rank 129 in a sample
of 183 countries. However, Brazil’s ranking fell slightly compared to its position in the previous
survey as the country failed to keep up with reforms in the rest of the world. Between 2015 and
2016, 137 economies worldwide implemented 283 business regulatory reforms. This represents an
increase of more than 20 percent compared to the previous year (Doing Business report, 2017).2
2. Brazil shows weaknesses across several categories of Ease of Doing Business

indicators. The procedures to start a business,
Brazil: Doing Business Indicators, 2017
obtain construction permits, pay taxes, and (Ranking)
trade across borders, are uncommonly Starting a Business
cumbersome. Slow progress on these issues
200
Dealing with
Resolving Insolvency
150 Construction Permits
over time has put Brazil below its peers in
100
terms of business’ attractiveness relative to Enforcing Contracts Getting Electricity
50
countries that have meanwhile implemented
0
reforms in many areas. Brazil does better in Trading across
Registering Property
comparisons related to getting electricity, Borders
enforcing contracts, resolving insolvency and

Paying Taxes Getting Credit
protecting minority rights.
Protecting Minority
Investors
1
Prepared by Izabela Karpowicz (WHD) and Kenji Moriyama (SPR).
2
In 2016, 137 economies worldwide implemented 283 business regulatory reforms. This represents an increase of
more than 20 percent compared to the previous year (Doing Business report, 2017).

BRAZIL
3. Looking across subcomponents of the overall score, Brazil’s global business

attractiveness improved in some areas but
is overall weaker than the LAC average. To Brazil: Doing Business Rankings
(Index from 0 to 100, where 0=lowest, 100=best)
facilitate comparison across time, made
difficult by changes in methodologies3, the Starting a business
100 Dealing with
DB database includes a measure called Resolving
Insolvency 80
Construction
“distance to frontier” (DTF) which measures

Permits
60
performance as the deviation of a country Enforcing Contracts
40 Getting Electricity
20
from the best performer. The DTF score
0
ranges from 0 to 100, with 0 representing
Trading Across Registering
the worst performer and 100 the frontier.4 Borders Property
Brazil’s DTF has been overall stable, with

some improvement in recent years, notably Paying Taxes Getting Credit
in the areas of insolvency procedures, Protecting Minority

DB2017
Investors DB2010
contract enforcement, and minority investor Sources: Doing Business, The World Bank. LAC-Brazil (Average)
protection. Starting a business is also easier

now, albeit still more complicated than elsewhere in the region. In a few areas, such as trading
across borders, Brazil moved away from the frontier. In terms of the overall business environment,
the largest five countries in the region (excluding Brazil), the LA5, are much closer to the frontier
than Brazil.
Brazil: Changing Distance to Frontier, 2010-17 Comparison

2017 2010-2017
Rank DF Below Latam Worsening Improvement
Overall 123 56.5 x x
Starting a business 175 65.0 x
Resolving insolvency 67 49.2
Dealing with construction permits 172 51.3 x
Paying taxes 181 33.0 x x
Trading across borders 149 55.6 x x
Getting electricity 47 81.2 x
Registering property 128 52.6 x
Enforcing contracts 37 67.4 x
Protecting minority investors 32 65.0 x
Getting credit 101 45.0
Source: Doing Business, The World Bank.
3 Methodological changes have been introduced most recently in 2011, 2014 and 2015 to expand coverage by
introducing new indicators and revising the components and methodologies of existing indicators.
4 The DTF score is calculated in two steps. First, DTF is calculated for each individual indicator using the formula: DTF
for country j, indicator k= (Worst Performance – Country j’s score on indicator k)/(Worst Performance—Best
Performance). Second, the DTF scores for each individual indicator are consolidated through simple averaging into
one DTF score for each of the ten topics, which ultimately provides a country’s ranking on the topic. The country’s
overall ranking is a simple average of its rank across all ten topics.

BRAZIL
Brazil: Ease of Doing Business–Distance to Frontier

(Index, 0-100, higher is better)
80
Mexico Peru Uruguay Brazil Chile Colombia
75
70
65
60
55
50
2010 2011 2012 2013 2014 2015 2016 2017
Sources: Doing Business, The World Bank.
4. Ease of doing business in Brazil is low relative to Brazil’s income. Plotted against per
capita GDP in 2015, Brazil’s DTF appears too low (Brazil’s index would have to move from the current
56 to 65 percent to approach the fitted line. Such an improvement would be difficult to achieve over
a short period. In the latest DB, Kazakhstan was among the countries that increased the most its DTF
in one year owing to reforms across seven dimensions of the index; still, the improvement was less
than 5 percentage points.
5. What makes Brazil’s position weak?
 Starting a business is a cumbersome process Brazil: Doing Business and Per Capita GDP
(In logs)
in Brazil. An estimated 11 procedures (down 12
from 14 in 2010) and about 80 hours are 11

Per Capita GDP (logs)
necessary to start a business. Although some 10
improvement was achieved over the years in 9 Brazil
8
this dimension, Brazil scores relatively low
7
compared to the average in Latin America. 6
5
 Paying taxes. It takes 2038 hours a year for a 4
firm to pay taxes in Brazil, an absolute 3.5 3.7 3.9 4.1 4.3 4.5
maximum in the sample. This number fell

Distance to Frontier (logs)
Sources: DOB, WDI; and IMF staff estimates.
from 2600 in 2010, a large improvement, but
still insufficient considering this is double the time it takes to pay taxes in the second less
efficient country according to this measure in Latin America, Bolivia. The reduction compliance
time in Brazil was the result of electronic systems that were introduced some years ago resulting
in more efficient tax compliance processes. However, although the time to pay taxes fell, the
share of total taxes and contributions in pre-tax profits increased and it is estimated at slightly

BRAZIL
below 70 percent. The incidence of taxes in profits is only 30 percent in Chile, for instance.5
OECD (2015) points to the fragmented and inefficient indirect tax system as one of the
important contributors to the “Brazil cost.” Tax credits for intermediate inputs are allowed only
when they are embodied in the final good sold. The burden of proof regarding how much of an
input goes directly into the final products lies with taxpayers, resulting in extensive use of tax
accountants and frequent lawsuits over disputes.
 Tax rates and tax regulations, alongside corruption and inefficient bureaucracy, have been
identified as the most problematic factors for doing business in Brazil in the 2015 World
Economic Forum’s survey.6
Brazil: Most Problematic Factors for Doing Business (2015)
 Among its political risk 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
index components, the Tax rates
International Country Corruption

Tax regulations
Risk Guide (ICRG) Inefficient government bureaucracy
Policy instability
shows a low score—2 Restrictive labor regulations
Inadequate supply of infrastructure
out of 6—on Inadequately educated workforce
corruption (indicating Access to financing

Government instability
high level of “potential Insufficient capacity to innovate
Inflation
corruption in the form Poor public health
Poor work ethic in national labor force
of excessive Crime and theft
patronage, nepotism, Foreign currency regulations
job reservations,
Source: World Economic Forum's Executive Opinion Survey.
‘favor-for favors’, secret party funding, and suspiciously close ties between politics and
business”) and a low score on the quality of bureaucracy—2 out of 4—intended as freedom from
political interference and interruption of government services. While these issues affect the
Bureaucracy Quality Corruption

(Index: max=4) (Index: max=6)
4.0 6.0
5.0
3.0
4.0
2.0 3.0
2.0
1.0
1.0
0.0 0.0
Brazil Latin America Emerging OECD Brazil Latin America Emerging OECD
& Caribbean Markets & Caribbean Markets
Source: International Country Risk Guide.
5 In the sample, it takes on average 251 hours to comply with taxes, there are 25 payments, and the total tax rate is
40.6 percent. (World Bank, 2016, Paying Taxes 2017: The Global Picture, World Bank Group and PwC.) Taxes and
contributions include: the profit or corporate income tax, social contributions and labor taxes paid by the employer,
property taxes, property transfer taxes, dividend tax, capital gains tax, financial transactions tax, waste collection
taxes, vehicle and road taxes, and any other small taxes or fees.
6
From the list of factors, respondents to the World Economic Forum's Executive Opinion Survey were asked to select
the five most problematic factors for doing business in their country and to rank them between 1 (most problematic)
and 5. The score corresponds to the responses weighted according to their rankings.

BRAZIL
operation of the government more directly, they tend to have repercussion on the business
environment given that the public sector is a large customer, especially in the services industry
in Brazil. Moreover, low trust in policies and corruption can make private sector less forward-
looking and hamper accumulation of physical as well as human capital, constraining productivity
gains.
Institution Related GCI Scores (2016-17) Institution Related GCI Scores (2016-17)
Part I Part II
Property rights
Transparency of
Efficiency of legal 6
Intellectual property government policymaking
framework in 6
5 protection
challenging regs.
4 Strength of investor 5
Organized crime
Efficiency of legal protection 4
3 Diversion of public
framework in settling 3
2 funds
disputes
2
1
1
0 Protection of minority Reliability of police
Burden of government Public trust in shareholders’ interests 0 services
regulation politicians
Wastefulness of Irregular payments Efficacy of corporate

government spending and bribes Ethical behavior of firms
boards
Favoritism in
decisions of Judicial independence Brazil Brazil
Strength of auditing and
government officials LATAM LATAM
reporting standards
Major EMs Major EMs
Source: World Economic Forum.
Source: World Economic Forum. AMs AMs
 Trading across borders. Brazil occupies a very low rank of 149 in this indicator. DB records the
time and cost associated with the logistical process of exporting and importing goods.7 Brazil’s
border compliance and import/export costs are among the highest in the world. It takes more
than 5 days to acquire customs clearance and inspections for exporting from Brazil, and more
than 15 days to clear imports, which is slower than for most of the countries in the world.
 Getting credit. Brazil occupies rank 101 in this indicator. Credit information is deep and Brazil’s
credit bureau coverage per adult is high, but the credit registry can be strengthened. Legal
rights are also an area that calls for improvement.
 Insolvency. The insolvency framework has improved, but recovery rates below 16 percent of the
value, are particularly low, and time to conclude the process (4 years on average) is long.
B. Recent Reforms Affecting the DB Score
6. Brazil’s business environment has changed over the past decade. While many changes
were positive, some worsened its position in international rankings. Some important changes in the
policy environment, such as the expanded presence of public banks and the increased state
intervention in the electricity and other key sectors, however, are not necessarily reflected in the list
below, which corresponds to the standard criteria used in DB.
7DB measures the time and cost (excluding tariffs) associated with three sets of procedures—documentary
compliance, border compliance and domestic transport—within the overall process of exporting or importing a
shipment of goods.

BRAZIL
 Starting a business. In 2016, an online portal for processing business licenses was launched in
Rio de Janeiro; although the opening hours of the Rio de Janeiro business registry were reduced,
the overall effect should be positive. Improvements in the synchronization between federal and
state-level tax authorities took place between 2009 and 2010.
 Paying taxes. Paying taxes became less costly for companies with the repeal of the tax on check
transactions in 2011.
 Trading across borders. Over the past few years trading across borders has become easier
thanks to an electronic system that reduced the time required for imports documentary
compliance. Time for documentary and border compliance for exporting was also reduced by
implementing the electronic SISCOMEX Portal system. These reforms came after a long period
without upgrades as the last improvement in the electronic data interchange system before that
took place in 2009.
 Getting credit. Private credit bureaus have been allowed to collect and share positive
information since 2011.
 Enforcing contracts. Contract enforcement has been strengthened thanks to a new mediation
law—that includes financial incentives for parties to attempt mediation—and a new code of civil
procedure. In 2012, an electronic system for filing initial complaints was established at the São
Paulo civil district court.
 Registering property. In 2012, transferring property became more difficult with the
requirement of a new certificate on good standing on labor debts, adding to the number of due
diligence procedures. Moreover, in 2015, in the city of São Paulo increased the property transfer
tax.
C. Announced Micro Reform Measures

7. The government has recently announced a package of micro measures seeking to
improve the business climate by streamlining administration and reducing red tape. While
some of these measures will not be captured directly in the DB indicators, they have the potential to
improve economic efficiency and facilitate the opening of the country. The timeline for their
implementation is short, but the effects may take some time to materialize. The package includes
47 actions directly linked to the Ministry of Industry, Foreign Trade and Services (MDIC), or to
agencies under the ministry’s umbrella. Many measures—24 out of 47—are linked to foreign trade
and consist of digitalization, computerization, and standardization of procedures.
 Starting a business. A National Network for Registration Simplification and Leading of Business
(Redesim) will accelerate opening of new companies by integrating cadasters with registries and
licensing bodies.
 Paying (and collecting) taxes. The government aims at simplifying procedures for tax refunds
and the payment of labor and tax obligations. To that purpose an electronic collection system
will be established (eSocial). The system currently only applies to taxes and contributions

BRAZIL
collected from domestic workers and will be extended to unify 13 different obligations.
Electronic invoicing used by the states and municipalities, and accounting information provided
to the Federal Revenue will be simplified and standardized.8 Moreover, the procedures
undertaken by businesses to declare and submit tax-related documents will be streamlined. The
reform of the overly complex Programa de Integração Social (PIS)9 will simplify the definition of
the tax base and thus reduce the currently high volume of litigation demands.
 Trading across borders. The agenda includes numerous measures to streamline the steps
involved in import/export activities. A Single Foreign Trade Portal accessible via internet will
consolidate the required documents including digital certificates of origin, simplify trade
processes and customs transit, lowering the number of days necessary to export (from 13 to 8
days) and import (from 17 to 10 days). Other measures range from simple changes—such as
ending the need to register to access the MDIC’s official statistics online—to issues that affect
the complex regime which grants exemption from import tariffs to exporters. Accreditation
procedures for laboratories and organizations will be reviewed (Inmetro) and patent and brand
registration procedures will also be standardized. Important efforts will be made to cut the
waiting time for reviewing the annual reports of research and development projects which grant
access to some tax breaks in the Manaus Tax Free Zone.
 Labor productivity. Recently approved in Congress, the Migration Law allows concession of
working visas for citizens of other countries who are in possession of higher education diplomas.
This measure may open the labor market to competition and could alleviate to some extent the
pressures on the demand for skills in Brazil.
 Getting credit. Improvements will affect the positive credit registry and reduce informational
asymmetry; a centralized market for electronic credit card and other receivables will be created;
legislation for the fiduciary alienation of immovable property that goes to auction will be
strengthened. Improvements of the Bankruptcy Law aim to empower creditors, promote
extrajudicial recovery, streamline processes, and strengthen the rights of the acquirers of the
company.
 Other measures. The government has committed to implementing a broader set of measures
affecting the labor market, the payment system, and some credit allocation mechanisms. An
attempt is being made to coordinate a settlement of tax and social-security debts of companies;
large companies’ debts to the Brazilian Development Bank (BNDES) under the Investment
Support Program (PSI) are under renegotiation to support deleveraging and step up growth on
new credit and activity; small and medium-sized companies’ debts are also being renegotiated.
Other micro reforms include authorization for merchants to charge different prices for purchases
8 This has already been implemented in São Paulo, Fortaleza and Porto Alegre municipalities.
9 PIS is the contribution paid by companies to finance unemployment insurance and dismissal of employees who
earn up to two minimum wages.

BRAZIL
paid in cash or with card; the reduction in the period for repayment and in the cost of revolving
credit to the consumer, and the standardization of forms for commercial establishments.
D. External Competitiveness
8. Brazil’s external sector performance has deteriorated in recent years.10 Since the global
financial crisis, investment, export growth and market efficiency have declined relative to peers. High
unit labor costs, inadequate infrastructure (IMF Working Paper No. 15/180), elevated tariffs, and
high rigidities in factor, goods, and financial markets have reduced openness and compressed
Brazilian exporters’ shares in global markets. Export volumes performance no longer surpasses
peers. Brazil witnessed strong export performance up to the global financial crisis,11 in the upper
quartile of major EMs.12 However, following the crisis, its volume path has converged to the median
of major EMs. This reflects, inter alia, ample swings in the REER during the 2000s―a large
depreciation in the first half and appreciation thereafter, in sharp contrast with other EMs and
LATAM countries. (That said, export volumes in some EMs, especially in Asia, may be inflated by re-
exports of goods because of their integration into global value chains.)
Export Volume (2000-15) Terms of Trade and REER (2000-16)

Goods and Services. Index: 2000=100 Emerging markets. Index: 2000 = 100
300
Major EMs (25-75 percentile) 160
280 LATAM 2010
Brazil
260 Major EMs (Median)
150 AMs
240 AMs (Median)
LATAM Major EMs
220
200 140 Brazil
180
160 130
140
REER
120 120
100
110
Source: WEO.
100
90
802003
90 100 110 120 130 140 150 160 170 180
Terms of Trade
Sources: INS and WEO.
10
Canuto et al. (2013) and OECD (2015).
11 This is a robust observation even with the trade volumes are normalized as 1995=100.
12 The sample includes: Argentina, Chile, China, Colombia, Egypt, Hungary, India, Indonesia, Malaysia, Mexico,
Morocco, Peru, Philippines, Poland, Romania, Russia, Thailand, Tunisia, Turkey and South Africa. AMs follow the WEO
classification.

BRAZIL
Export Diversification Exports: Change in Domestic Value Added

(HH Index) Share in Major EMs
Change from 2000 to 2011, percentage points

0.16 15
Major EMs (25-75 percentile)
Brazil PHL
0.14 10 CHN
LATAM MYS
0.12 Major EMs (Median) 5 IDN BRA
AMs HUN MEX CHLRUS
0.10 0 ROM
0.08 COL
-5 TUN ZAF
0.06 THA
-10 POL ARG
0.04 y = -0.367x + 21.937
-15 R² = 0.3385 TUR
0.02
-20
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
IND
Other EMs Commodity exporters (WEO)
-25
Source: WITS.
30 40 50 60 70 80 90 100
2000, percent
Sources: OECD and WTO.
9. Brazil’s exports are diversified,13 but this advantage appears to have diminished
recently. The Herfindahl concentration index was in line with the EM median in 2014‒15, up from
lower levels registered over the past two decades. The domestic content of Brazil’s exports, based
on value added trade data compiled by the OECD and WTO, has been persistently higher than its
peers’ pointing to the weight of commodities in Brazilian exports and the country’s lower integration
into global value chains.14 Brazil’s export quality, compiled by Henn et al. (2015),15 is well below
average implied by the frontier of other major EMs and advanced economies (AMs).16
E. Factors Explaining Deteriorating Competitiveness
While aggregated macro indicators, such as export volumes, are common summary
statistics, they may be affected by numerous factors. This suggests the analysis needs to
be complemented by micro indicators.
10. As shown in other studies, trade liberalization appears to have stalled in Brazil after
the large reduction in tariffs in the early 1990s. Tariffs in Brazil are higher than in its peers,
despite efforts to liberalize trade since the early 1990s (Box 1). While import tariffs in Brazil have
declined since the early 1990s to below 10 percent, Brazil’s tariff regime is currently more
protectionist than the average in Latin America and the Caribbean (LAC) and the upper middle
income country average. Disaggregated data suggests that goods and other raw materials tariffs are
substantially higher than Brazil’s peers’, especially on capital goods.
13
Ding and Hadzi-Vaskov (2017) report a similar observation.
14
Lower integration into global supply chains could explain part of the recent export volume developments in
Brazil.
15
Henn et al. (2015) also reported that institutional quality, liberal trade policies, FDI inflows, and human capital
could promote quality upgrading.
16
Commodity exporters generally display higher domestic content in exports and lower export quality.

BRAZIL
Box 1. History of Trade Liberalization in Brazil

By the late 1980s, Brazil's import policy exhibited the following basic characteristics (Cardoso 2009):
 Tariff structure was still largely the same as in 1957.
 Generalized presence of tariffs with redundant quotas.
 Various additional taxes, such as financial transaction tax, port improvement tax, and additional tax for
renewal of maritime transportation.
 Ample use of Non-Tariff Barriers (NTBs), such as a list of products with suspended import licenses,
specific advance authorization for certain products (steel and IT products) and annual corporate import
quotas. Cardoso (2009) estimated that tobacco, real estate, plastic products, clothing, footwear and
textile products, perfumes, soap and candles, and transport material were most protected by the
barriers in 1987.
 The existence of 42 special regimes, allowing for the exemption or reduction of taxes.
Liberalization occurred during the
early 1990s. Trade liberalization, Brazil, Tariff Rates (1990-Latest)
including the three tariff reduction Weighted average, percent
waves, was implemented during 1988 35
1990 2000 2010 Latest (2015)
and 1994. In the first phase (1998‒89), 30
two tariff reforms were undertaken 25
seeking to eliminate the redundant
20
share of the nominal tariff without
15
significantly impacting import volumes.
The second phase (1991‒93), gradual 10
import tariff reductions, was 5
implemented after NTBs were abolished 0
in 1990, with imports being controlled All Products Capital goods Consumer Intermediate Raw materials
through import tariffs. The third phase goods goods
Source: World Integrate Trade Solution (WITS).
(1994) intended to discipline domestic
prices through greater external competition. In response to these measure, and alongside measures taken in
1994 to facilitate the implementation of a stabilization plan (the Real Plan), the weighted-average tariff
declined substantially.
But there was a partial reversal thereafter. The liberalization during the first half of the 1990s, however,
soon gave way to a reversal as Brazil entered a period of increasing current account deficits driven by a
severe exchange rate appreciation and negative external shocks (Moreira 2009). In November of 1997, the
government temporarily raised the tariffs by 3 percentage points. After that, the weighted average of tariff
rates demonstrated a very gradual decline until the global financial crisis, and gradually increased since then.
De Araujo Jr. (2017) pointed that frequent use of anti-dumping, especially after 2008, has contributed to
eroding competitiveness by keeping domestic prices high and reducing competition in domestic markets.
The literature identifies a positive link between trade liberalization and productivity. Lopez-Córdova
and Moreira (2004) reported that the trade liberalization of the first half of the 1990s made a substantial
contribution to lowering prices of capital goods in Brazil, estimating that a 10 percent reduction in tariffs
increases total factor productivity (TFP) by 1 percent. Lisboa et al. (2010) used firm-level data and found that
the reductions of tariffs on inputs were important in explaining productivity growth during trade
liberalization in Brazil. Tariff reduction in products that are mainly used as inputs affect productivity growth.

BRAZIL
11. Investment in Brazil has been weaker than in its peers. Investment shows some
correlation to tariffs on capital goods across countries. Because weak investment relative to peers
was persistent in Brazil, the quality of its capital stock has deteriorated. The slow growth and the
aging of the capital stock has constrained marginal productivity of labor, limiting its growth
potential.
Figure 1. Brazil: Investment and Tariffs: Brazil and Emerging Markets

Brazil
Brazil and Major EMs, All Products, Tariff Rates, 1996-Latest Tariffs across EMs, Latest
Weighted average, percent, median LATAM
Weighted average
Major EMs (75 percentile)
25
All products Major EMs (median)
1996 2000 2005 2010 Latest 15
20
10
15 Raw materials 5 Capital goods
10 0
Intermediate
0
Consumer goods
goods
Brazil Asia Europe LATAM Middle East
Source: World Integrate Trade Solution (WITS). and Africa
Source: World Integrate Trade Solution (WITS).
Gross Fixed Capital Formation (2000-16) Tariffs on Capital Goods and Investment (2011-15)
In percent of GDP Major EMs, excluding China
30
Major EMs (25-75 percentile) 35
Investment/GDP
Brazil
Major EMs (Median) IDN IND
AMs (Median) 30 MAR
25
LATAM MYS
PER
25 THA
COL
CHL
20 MEX TUN
20 RUS BRA
TUR
ZAF
ARG
15 PHL
15 EGY y = -0.4022x + 24.661
R² = 0.0498
10
Source: WEO. 0 2 4 6 8 10 12
Tariffs on capital goods (%)
Sources: WITS, WEO.
Major EMs. Vintage of Capital Stock, 1990-2021 Capital Stock Vintage and ICOR (2010-16)
Median Major EMs, excluding China
0.65
ICOR
Vintage, years
Major EMs, 25-75 percentile

18 More investment is needed to
Europe 0.60 HUN
LATAM support growth
0.55
16 Asia
ARG
Major EMs (21 countries) 0.50
PER
Middle East and Africa MAR
14 0.45 CHL PHL BRA
0.40 ROM
TUR THA TUN RUS
12
0.35 POL COL
IDN ZAF
0.30 EGY
10 MEX
MYS
0.25 y = 0.0158x + 0.2
R² = 0.1635
IND
8 0.20
6 8 10 12 14 16 18
Capital stock vintage, years
Sources: PWT 9.0 and WEO.
Sources: WEO and PWT

BRAZIL
12. Non-tariff measures FDI Restrictiveness Index (1997-2015)

are used extensively in Brazil. Higher number indicates more restrictive, average
Quotas and special safeguards 0.6
are more common in Brazil than
0.5
in LAC countries on average
1997 2006 2015
(Cardoso 2009). OECD (2014) 0.4
finds that Brazil is second only to
0.3
Indonesia in the number of local
content requirements imposed 0.2
since the global financial crisis. 0.1

However, restrictions to inward
0.0
FDI are close to peers. Brazil has
Brazil LATAM Asia Europe Middle East
fewer restrictions on FDI than and Africa
Source: OECD.
Asian and European EMs,
although its relative advantage appears having diminished recently. FDI has a role not only for the
financing of current account deficits but also for transferring technology/stimulating productivity
gains.
Figure 2. Global Competitiveness: Brazil and Emerging Markets
Labor Related GCI Scores, 2016-17 Brazil Goods Markets Related GCI Scores (2016-17) Brazil
Cooperation in LATAM LATAM
labor-employer Major EMs
Major EMs
relations Intensity of local AMs
6 AMs competition
5 6 Extent of market
Buyer sophistication
Capacity to attract 5 dominance
4 Hiring and firing
talent 4
3 Degree of customer Effectiveness of anti-
orientation 3 monopoly policy
2
2
1 1
0 Burden of customs Effect of taxation on
procedures 0 incentives to invest
Capacity to retain
Flexibility of wage
talent
Business impact of No. of procedures to
rules on FDI start a business
Prevalence of foreign No.of days to start

Professional Pay and ownership business
management productivity
Prevalence of trade
Source: World Economic Forum. Source: World Economic Forum. barriers
Financial Sector Related GCI Scores (2016-17) Brazil

LATAM
Affordability of Major EMs
financial services AMs
6
Financial services
Legal rights index meeting business
4
needs
2
Regulation of Financing through
securities exchange 0 local equity markets
Soundness of Ease of access

banks loans
Venture capital
availability
Source: World Economic Forum.

BRAZIL
Box 2. Literature on the Impact of Trade Liberalization on Productivity and Growth

Theoretical literature points to several channels through which trade liberalization can boost
productivity and external competitiveness (Ahn et al. 2016).
 Competition. Lower trade and FDI barriers can strengthen competition in the liberalized sector(s),
by putting pressure on domestic producers to lower price margins, exploit economies of scale, improve
efficiency, absorb foreign technology, or innovate.
 Scale. Productivity gains from liberalization may accrue disproportionately to larger and more
productive firms, enabling them to gain market share and amplify aggregate gains within the liberalized
sector.
 Lower input costs. Trade liberalization, which includes local content requirements, can boost
productivity by increasing the quality and variety of intermediate inputs available to domestic producers.
But higher productivity does not necessarily imply sustained higher growth. As demonstrated by
Bajona et al. (2008), standard trade models—including Ricardian models, Heckscher-Ohlin models,
monopolistic competition models with homogeneous firms, and monopolistic competition models with
heterogeneous firms—at best predict that trade liberalization increases welfare, but does not necessarily
sustain real GDP growth. Trade liberalization improves resource allocation, but the real income gains may be
limited to the period during which trade liberalization occurs (Moreira 2004).
Sectoral and firm-specific factors are also important for growth. Mechanisms outside of those analyzed
in standard trade models can help liberalization lead sustained higher growth. Recent theoretical and
empirical evidence suggests that the impact of trade liberalization varies widely across firms, depending on
their individual characteristics, such as ownership structure (foreign owned vs. domestic), the extent to which
they use imported inputs, and the degree of competition in their industry (Ahn et al. 2016). Also, the
treatment of the nontraded sector is important in determining the magnitude of liberalization’s effect on
labor productivity (Kovak 2013). This suggests that interactions between trade liberalization and other
policies, such as product and labor market regulation or barriers to FDI that could affect firm characteristics
is important.
The degree of integration matters too. Also, the degree and quality of the integration, both in terms of
trade-to-GDP ratios and the level and structure of protection, tend to have a major influence in the end
results of trade liberalization (Moreira 2004). Gains from trade are maximized when a country not only opens
its own market, but when it also enjoys access to markets abroad (Moreira 2009). Monfort (2008)
summarizes that benefits from an FTA mostly stem from non-trade channels, including reduced risk
premiums, increased foreign direct investment, or improved factor productivity, rather directly from trade.
Beaton et al. (2017) employed a large panel data set to show that growth is positively affected by trade
openness, trade partner diversification, integration to global value chains, and export quality. Hannan (2017)
found that trade agreements in Latin America have generated substantial growth in exports, although the
export gains in Latin America were more limited than the world average, possibly reflecting the region’s
lower trade openness and weaker integration into global value chains.
Empirical, cross country studies have underscored gains in productivity and growth from trade
liberalization (Goldberg et al., 2004). Wacziarg and Welch (2003) employed an event study approach
using aggregate cross country data to find the robust positive effect of trade liberalization on growth. Ahn
et al. (2016) used data of effective tariffs in 18 sectors across 18 advanced countries spanning over two
decades and found a significant and robust impact of input tariff liberalization on sector-level total factor
productivity (TFP)— a one percentage reduction in input tariffs raises TFP levels by about two percent.
Krishna and Mitra (1998) investigated the effects on competition and productivity of the dramatic 1991 trade
liberalization in India. Using firm-level data, they found (i) strong evidence of an increase in competition and
(ii) some evidence of an increase in the growth rate of productivity.

BRAZIL
Box 2. Literature on the Impact of Trade Liberalization on Productivity and Growth

(Concluded)
Liberalization must be supported by other structural policies. Benefits of trade liberalization can be
enhanced through coordinated measures to improve non-trade factors such as foreign investment, rigidities
in labor and product markets, and global markets’ integration. OECD (2011) reported a substantial
improvement in real GDP from a coordinated reduction in trade costs related to tariff and non-tariff barriers,
but most of the improvement was driven by a decline in non-tariff barriers. Rahman et al. (2015) found a
strong role of structural reforms for taking advantage of the tariff-free trade environment to export in new
member states of EU. Especially, they reported that (i) policies that influence higher education and skills
match, incentives to work, and foreign investment environment are most relevant and (ii) reform sequencing
becomes important for export quality improvement.1 Also, Baniya (2017) found that timeliness of delivery of
goods constitutes a comparative advantage of exports. IMF (2015) also showed that despite the positive
effect of trade agreements on exports they need to be accompanied by structural reforms and reduction of
non-tariff barriers.
1
A conducive environment for foreign investment and greater links with supply chains are key for countries at the lower
end of quality spectrum, while tertiary education, skills upgrade, and R&D spending are priorities for countries at the
medium-level of quality spectrum.
F. Benefits of Furthering Trade Liberalization in Brazil
13. Trade protection should be reduced steadily by lowering tariffs, removing non-tariff
barriers and scaling back local content requirements, including on public procurement.
OECD’s extensive quantitative analysis on the impact of trade liberalization, as well as the work of
other analysts, points to substantial gains from liberalization, if appropriately implemented.
Lowering barriers to trade allows firms to use a higher share of foreign intermediate goods in
production. Final goods are thus sold at lower prices, enhancing the competitiveness of Brazilian
exports while also benefitting Brazilian households. In addition, lower barriers to trade reduce the
cost of capital, spurring investment and supporting further expansion of production.
 OECD (2011) presented a simulation derived from large scale global general equilibrium models
to assess the effects of the coordinated reduction in trade costs related to tariff and non-tariff
barriers. The main policy scenario assumes that G20 member countries implement a 50 percent
MFN tariff reduction and a reduction of non-tariff barriers by 50 percent on an MFN basis. The
estimated impact for Brazil is an increase in the real GDP by about 7½ percent over the long-
run, most of it driven by a decline in non-tariff barriers.
 Araujo and Flaig (2016) used a computable general equilibrium model to study the impact of
trade liberalization.17 The simulation assesses the impact of a reduction in import tariffs (to the
OECD minimum level), local content requirements (eliminate all requirements), and taxes levied
on exports (to zero) in Brazil. The results show that: (i) aggregated production level in Brazil
would increase by about 1.7 percent, driven by a pick-up in exports; (ii) lifting impediments to
17
This is a comparative static, constant returns to scale multi-region Computable General Equilibrium model that
captures inter-industry effects while tracking differences in trade patters by individual country and sector. For details,
see Box 2 in Araujo and Flaig (2016).

BRAZIL
trade would allow a deeper integration to global value chains, as implied by the substantial
growth in some manufacturing sectors and non-ferrous metal exports; and (iii) the largest gains
in production would stem from zero-tax rates on exports, while for exports the gains would
come from eliminating tariffs on imports of intermediate goods.
 IMF (2017) analyzes the impact on investment of tariffs levied on capital goods. The result
indicates that, if Brazil halves its average tariffs on capital goods (about 10 percent in 2015), it
would increase investment by about 10 percent over the medium-term. Halving tariffs on capital
goods in Brazil would raise its investment-to-GDP by about 2 percentage points. This is
equivalent to an increase in investment of about 15 percent over three years assuming a real
GDP growth of 2 percent on average.18
14. Greater integration into the global economy should improve external competitiveness.
Recent studies on LAC countries show significant gains in external competitiveness and potential
growth occurring through trade openness and participation to the global value chains.
 Beaton et al. (2017) and Cerra and Woldenmichael (2017) show a statistically significant positive
impact of trade integration (openness and integration to global value chains) to growth. Based
on their estimated parameters, the probability of export surge in Brazil could increase by about
8 percentage points, if the degree of globalization in Brazil, measured by the KOF index of
globalization,19 increases from the latest reported level to its frontier implied by per capita
income among AMs and major EMs. Also, raising the degree of participation in global value
chains from the current level (slightly below the 10th percentile) to the 25th and 50th percentiles
of AMs and major EMs would raise annualized growth by 0.15 percent and 0.3 percent,
respectively.
 Cerdeiro (2016) simulated the effects of the TPP in Latin American countries, both the TPP and
non-TPP member countries, using a multi-sector model. While the prospect of TPP is highly
uncertain, the study can still highlight the potential gains for Brazil by joining large “mega” trade
agreements. First, liberalization of tariffs and non-tariff barriers in TPP members would reduce
exports in percent of GDP by 0.2 percentage points. Second, expanding the TPP to include nine
Latin America and Caribbean (LAC) countries,20 including Brazil, would raise real income in Brazil
by 0.4 percent.21 Also, using the estimated parameters in Beaton et al. (2017), increasing the
18
This assumes constant investment and GDP deflators.
19
The index is defined over three dimensions: economic globalization, characterized as long distance flows of goods,
capital and services as well as information and perceptions that accompany market exchanges; political globalization,
characterized by a diffusion of government policies; and social globalization, expressed as the spread of ideas,
information, images and people. For details, see http://globalization.kof.ethz.ch/.
20
Argentina, Bolivia, Brazil, Colombia, El Salvador, Guatemala, Nicaragua, Paraguay and Uruguay.
21
The simulation assumes a full tariff liberalization and a decline in non-tariff barriers to the levels of those of TPP
members with less restrictive trade regimes.

BRAZIL
number of trade partners in agreements to the 25the and the median of AMs and major EMs
would raise annual growth by 0.5 percent and one percent, respectively.
15. Progress in expanding and diversifying bilateral, regional and multilateral trade
agreements will be key for further enhancing integration. Until recently, Brazil has, through
Mercosur, developed an extensive network of tariff agreements in the region. Trade agreements
cover goods, services, investment, trade facilitation, technical regulations and government
procurement. The network has been further extended to Cuba, Guyana and Saint Kitts and Nevis.
Diversification of trade agreements, especially with the countries of the Pacific basin (Peru, Chile and
Colombia) was also sought. The government intends to expand trade agreements with Mexico while
negotiations to conclude the Mercosur agreement with the EU, which accounts for 20 percent of
trade of the block, were resumed. Moreover, the bilateral agenda with the U.S., including the
implementation and expansion of the agreements on Regulatory Convergence and Trade
Facilitation, has been already signed.
16. Reforms in product and labor market regulation are critical for improving external
competitiveness. In this context, it would be recommended:
 strengthening competition by streamlining regulation on product markets and implementing

planned reductions in entry regulations;
 improving the technical capacity and planning for infrastructure concessions and elaborating
more detailed tender packages prior to launching tender calls;
 consolidating indirect taxes at the state and federal levels and working towards a value-added
tax with a broad base, full refunds for input and zero-rating of exports; and, more broadly,
 pursuing DB reforms.
G. Conclusions
17. Brazil can boost productivity through reforms aimed at improving the business
environment and external competitiveness. As argued in numerous studies, lowering tariff and
non-tariff barriers can enhance trade integration, improve economic efficiency, and boost
investment and growth. But pursuing trade agreements, over and above the Mercosur partnership,
bears promises of further global integration of Brazil and stronger growth. Product and labor market
regulation reforms are also important for strengthening channels through which the business
environment affects competitiveness and trade, as are reforms to the tax system. Implementation of
trade measure should, however, consider the cyclical position of the economy: tariffs reduction
could increase competition forcing some inefficient firms to exit from markets. Although this would
increase productivity over the long run, it could result in higher unemployment and lower
consumption in the short-run. This suggests that reforms should be carefully sequenced, paying
attention to their potential contractionary impact on economic activity as the economic
transformation unravels (IMF, 2016c).

BRAZIL
References
Ahn, JaeBin, Era Dabla-Norris, Romain Duval, Bingjie Hu and Lamin Njie, 2016, “Reassessing the
Productivity Gains from Trade Liberalization,” IMF Working Paper No. 16/77 (Washington:
International Monetary Fund).
Araujo, Sonia and Dorothee Flaig, 2016, “Quantifying the Effects of Lowering Barriers to Trade in
Brazil: A CGE Model Simulation,” OECD Economics Department Working Papers No. 1295.
Bajona, Claustre, Mark J. Gibson, Timothy J. Kehoe, and Kim J. Ruhl, 2008, “Trade Liberalization,
Growth, and Productivity,” Federal Reserve Bank of Minneapolis.
Baniya, Suprabha, 2017, “Effects of Timeliness on the Trade Pattern between Primary and Processed
Goods,” IMF Working Paper No. 17/44 (Washington: International Monetary Fund).
Beaton, Kimberly, Aliona Cebotari, and Andras Komaromi, 2017, “Revisiting the Link between Trade,
Growth and Inequality: Lessons for Latin America and the Caribbean, IMF Working Paper No. 17/46
(Washington: International Monetary Fund).
Canuto, Otaviano, Matheus Cavallari, and Jose Guilherme Reis, 2013, “Brazilian Exports. Climbing
Down a Competitiveness Cliff,” World Bank Policy Research Working Paper 6302 (Washington).
Cardoso, Eliana, 2009, “A Brief History of Trade Policies in Brazil: From ISI, Export Promotion and
Import Liberalization to Multilateral and Regional Agreements,” Paper prepared for the conference
on “The Political Economy of Trade Policy in the BRICS,” March 27-28, New Orleans, LA.
Cerdeiro, Diego A., 2016, “Estimating the Effects of the Trans-Pacific Partnership (TPP) on Latin
America and the Caribbean (LAC),” IMF Working Paper No. 16/101 (Washington: International
Monetary Fund).
Cerra, Valerie and Martha Tesfaye Woldemichael, 2017, “Launching Export Accelerations in Latin
America and the World,” IMF Working Paper No. 17/43 (Washington: International Monetary Fund).
de Araujo Jr., Jose Tavares, 2017, “Anatomia da prodecao antidumping no Brazil,” Textos Cindes
No. 45.
Ding, Xiaodan and Metodij Hadzi-Vaskov, 2017, “Composition of Trade in Latin America and the
Caribbean,” IMF Working Paper No. 17/42 (Washington: International Monetary Fund).
Goldberg, Pinelopi Koujianou and Nina Pavcnik, 2004, “Trade, Inequality, and Poverty: What Do We
Know? Evidence from Recent Trade Liberalization Episodes in Developing Countries,” NBER Working
Paper No. 10593 (Cambridge, Massachusetts: National Bureau of Economic Research).
Hannan, 2017, “The Impact of Trade Agreements in Latin America using the Synthetic Control
Method,” IMF Working Paper No. 17/45 (Washington: International Monetary Fund).
Kovak, Brian K., 2013, “Regional Effects of Trade Reform: What is the Correct Measure of
Liberalization?” American Economic Review, Vol. 103(5), pp. 1960‒76.
BRAZIL
Krishna, Pravin and Devashish Mitra, 1998, “Trade Liberalization, Market Discipline and Productivity
Growth: New Evidence from India,” Journal of Development Economics, Vol. 56, pp. 447–62.
Lisboa, Marcos B., Naercio A. Menezes Filhoz, Adriana Schor, 2010, “The Effects of Trade
Liberalization on Productivity Growth in Brazil: Competition or Technology?” RBE Rio de Janeiro,
Vol. 64, No. 3, pp. 277–89 (Jul.‒Set.).
International Monetary Fund, 2015, “Trade Integration in Latin America and the Caribbean: Hype,
Hope, and Reality,” Regional Economic Outlook: Western Hemisphere, October.
International Monetary Fund, 2016a, Brazil: 2016 Article IV Consultation Staff Report, IMF Country
Report No. 16/348 (Washington).
International Monetary Fund, 2016b, World Economic Outlook, October (Washington).
International Monetary Fund, 2016c, World Economic Outlook, Spring (Washington).
International Monetary Fund, 2017, “Trade Integration in Latin America and the Caribbean,” Western
Hemisphere Department, Regional Cluster Report, March.
Monfort, Brieuc, 2008, “Chile: Trade Performance, Trade Liberalization, and Competitiveness,” IMF
Working Paper No. 08/128 (Washington: International Monetary Fund).
Moreira, Mauricio Mesquita, 2004, “Brazil’s Trade Liberalization and Growth: Has it Failed?” IDB
Occasional Paper No. 24 (Washington).
Moreira, Mauricio Mesquita, 2009, “Brazil’s Trade Policy: Old and New Issues,” IDB Working Paper
Series, # IDB-WP-139.
OECD, 2011, “The Impact of Trade Liberalization on Jobs and Growth: Technical Note,” OECD Trade
Policy Papers, No. 107.
OECD, 2014, “Emerging Policy Issues: Localization Barriers to Trade,” TAD/TC Working Paper Mo. 17.
OECD, 2015, OECD Economic Surveys Brazil, November 2015, OECD.
Rahman, Jesmin, Ara Stepanyan, Jessie Yang and Li Zeng, 2015, “Exports in a Tariff-Free
Environment: What Structural Reforms Matter? Evidence from the European Union Single Market,”
IMF Working Paper No. 15/187 (Washington: International Monetary Fund).
Tiffin, Andrew, 2014, “European Productivity, Innovation and Competitiveness: The Case of Italy,” IMF
Wacziarg, Romain and Karen Horn Welch, 2003, “Trade Liberalization and Growth: New Evidence,”
NBER Working Paper No. 10152 (Cambridge, Massachusetts: National Bureau of Economic
Research).

BRAZIL
DISTRIBUTIONAL EFFECTS OF BRAZIL’S PENSION

REFORM1
A. The Reform of Social Security
1. Brazil social security system faces large current and growing future financial
imbalances absent reforms. With increasing average benefits relative to contributing salaries,
and a rapidly aging population, Brazil’s social security system cannot be sustained in the future
without reform. Pension spending competes with productive outlays, and, at current trends,
may hinder government’s ability to provide other public goods.2
2. The Congress is currently debating a social security reform. This reform addresses
many of the main sources of financial imbalance under the present system. The proposal
introduces a minimum retirement age for women and men, with some transitional
arrangements, extends the minimum contribution period, and adjusts the benefit payable at
retirement. In addition, the reform would over time reduce differences in the regimes for
private and public sector employees. The reforms would also limit the scope for the
accumulation of benefits, such as a retirement benefit and a survivors’ pension, and lower the
rural pension subsidy (see Appendix I for more detail).
3. Taking into account the transition to the new system, the reform will affect
formal workers who are employed currently, especially those who are farther away from
retirement. By bringing closer the various parameters of the private and public pensions
systems, the reform will do much to narrow inequalities over time. But since the reform
Pension Expenditure
18
(2014 or latest avalable estimate) RGPS Financial Results
(Percent of GDP)
16 1.0 1.0
Expenditure on pensions, in percent of GDP
14 Brazil 0.5 0.5
12 0.0 0.0
-0.5 -0.5
10
-1.0 -1.0
8
-1.5 -1.5
6 -2.0 -2.0
4 -2.5 -2.5
2 -3.0 -3.0
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
0
Urban Rural Total
0 5 10 15 20 25 30
Share of population age 65 and older, in percent of total
Sources: World Development Indicators; and IMF Staff estimates based on a sample of about 100 countries.
1 Prepared by Izabela Karpowicz (WHD), Vivian Malta, and Marina Mendes Tavares (SPR).
2See Cuevas and others (2017) for a more comprehensive description of the social security system and a
discussion on pension system sustainability. Brazilian usage typically requires calling retirement programs
“aposentadorias” while reserving the term “pensões” for other specific benefits, such as a survivor’s or a
disability benefit. In this chapter we will use the terms “pensions” and “social security” interchangeably, and
encompassing all of these various social security programs, unless explicitly indicated otherwise.

BRAZIL
modifies various parameters in the private system too, including contributions, retirement
incentives and the value of benefits, it will affect not only the pension system’s sustainability,
but also the distribution of income across various categories of workers.
 The minimum retirement age for men and women and higher minimum contribution time:
postponing retirement implies working and earning wages longer, thus contributing to the
pension system for longer and receiving retirement benefits for a shorter time. Typically,
however, in a defined-benefit system, benefits claimed at the time of retirement would be
higher because they would be calculated based on a salary history which includes the
additional time earning (higher) wages. However, longer required years of contribution can
create disincentives to participate in formal work for workers with spotty careers, who are
typically low-wage earners, because they may be able to qualify for non-contributory
pensions before they can complete the (longer) minimum contribution period.
 A new benefit formula: in the reform under consideration, replacement rates would decline
for all individuals who contribute for less than 40 years, which is the time it takes to claim a
full pension under the proposed new rules. Changes in the formula would have a relatively
smaller impact on pensioners who currently retire based on contribution-time alone
because these individuals tend to have substantially long contribution histories anyway.
The changes do not affect individuals who would receive the minimum pension, which is
equal to the minimum wage.
 Tightened access to rural pensions and BPC: increasing the minimum retirement age for
non-contributory pensions could have a negative impact on lifetime earnings of rural
workers and urban workers who receive the BPC as they will be receiving the minimum
pension (as before), but for a shorter time.
 Limited cumulation of benefits (survivors’ pensions): limiting the accumulation of benefits

would lower beneficiaries’ total incomes; however, a third of survivors’ pensions goes to
the 10 percent richest families.
B. Stylized Facts on the Social Security System and Inequality
4. At present, there are large differences in benefits received by private sector and
public sector employees. Public sector workers enjoy higher wages during their productive
life, at all points in their career and across most professions; the public-private wage
differential seems to be around 50 percent based on the 2014 PNAD (see Chapter 4).3
Moreover, public sector workers generally retire earlier than private sector employees under
existing rules, have higher average post-retirement incomes, and their retirement benefits are
often subject to larger annual revisions than those of their counterparts in the private sector.
3
Pesquisa National por Amostra de Domicílios (PNAD).

BRAZIL
Median Monthly Earnings, by Age Minimum wage Distribution of Earnings 1/

(Percent)
(In reais of 2015) 100%
4,000 90%
80%
3,000 70%
60%
2,000 50%
40%
30%
1,000
20%
10%
0 0%
20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 1 2&3 4 5 6 7 8 9 10
Income Deciles
Urban Private Sector Rural Private Sector
Public Sector Urban Private Sector Rural Private Sector Public Sector Retirees & Survivors 2/
Sources: PNAD; and IMF staff calculations.

Sources: PNAD. 1/ All earnings (from labor or pension) of workers (formal or informal) ages 18 and above.
2/ Individuals receiving pensions (those who work and those who are retired).
5. The pension system seems to reinforce other sources of inequality. Pension

benefits are unequal in part because the option to retire by length of contribution is most
easily exercised by earners with stable formal sector jobs. Moreover, many high-income
workers retire by time of contribution at a relatively early age, and then continue to work in
the private sector while they draw a pension. The highest net benefit (public transfer), defined
as the difference between total benefits received post-retirement and total contributions paid
during a person’s working life, accrues to richer individuals—public workers and high skilled
private workers—under the current system. In fact, the World Bank (2017) estimates that
around half of all pension “subsidies” country-wide accrues to the top income quintile of the
population, while only 4 percent accrues to the bottom 20 percent.
RGPS Retirement Benefits, 2016

(Average monthly reais)
2,500 Average Monthly Earnings (Age 55-65), 2015
2,000 (In reais)
10,000 8,999
1,500 Ave. National Income
1,000 8,000
Min. Wage 6,448
500 6,000
3,898
0 4,000 81% 2,661
2,002 1,539
Total Urban Rural 2,000
Age Invalidity Length of service 0
Sources: Previdê ncia Social; and IBGE. All Skilled Unskilled
Private urban workers Retirees who work
6. Inequalities persist also within the RGPS at the urban/rural level. Skilled workers in
urban areas have sufficiently long contribution histories to retire by contribution time,
especially men. In rural areas, workers, in particular women, retire according to age (men aged
60 and women aged 55 who have completed at least 180 months of work in family
agriculture) and receive a benefit equal to one minimum wage, which we call hereafter
“rural pension”. The number of beneficiaries of rural pensions is higher than the rural

BRAZIL
population above 55 per IBGE estimates, across all states, suggesting that there may be abuse
in the system.4
Share of Total Monthly Benefits by Income Level, 2017

Urban Rural
100
100
90
90
80
80
Age Contribution time Non-contributory 70 Age Contribution time Non-contributory
70
60
60
50
50
40
40
30
30
20
20
10
10
0
0
Source: Previdência Social.
RGPS Social Security Benefits, 2017
Rural Urban
Age Contribution time Age Contribution time
Non-contributory Non-contributory
100.0 100
99.5 90
99.0 80
98.5 70
98.0 60
97.5 50
97.0 40
96.5 30
96.0 20
95.5 10
95.0 0
MEN WOMEN MEN WOMEN
Source: Previdência Social.
7. Some benefits with a social assistance component could be better targeted. The
BPC-LOAS are pension-like assistance benefits available to those who do not qualify for a
retirement benefit based on the two conditions mentioned above. These assist individuals
who are 65 and above or disabled and whose household income per capita is below ½ of
the minimum wage. They receive an amount equal to the minimum wage and their
conditions are reviewed every two years. Due to possible duplication and recourse to
jurisprudence, this benefit is not well targeted to low-income families: around 40 percent
of these benefits accrue to the top two quintiles of the population and over 10 percent to
4
Intituto Brasileiro de Geografia e Estatistica (IBGE).

BRAZIL
the bottom two deciles (World Bank, 2017). The reform proposes to increase the minimum
age for obtaining the benefit.
C. Some Distributive Effects of Pension Reform
8. How will the pension reform affect gross and net incomes of contributors? We
analyze the effects of the proposed changes in the social security on the “lifetime incomes” of
various “representative agents” or “ideal types” (defined here to include their earnings from
work, retirement benefits, and any pensions they leave to their survivors), and the net lifetime
transfers they receive from (or pay to) the social security system. Each agent is a different type
of worker (and future retiree), whose lifetime earnings trajectory is constructed with data from
Brazil’s continuous households survey (PNAD, 2015). We show that in the new steady state,
after the transition period is over, the proposed reform will affect lifetime incomes of private
unskilled, public and rural workers only marginally while skilled urban workers will earn more
because they will be required to work longer.,5 Nevertheless, the reform has an important
component of progressivity: it would reduce the high net transfers that currently flow to the
higher income workers (and their survivors).
9. Below, we show the estimated lifetime income path for four representative male
workers. We consider a median worker who dies at age 80 and leaves a survivor’s pension to
his spouse (but no other dependents), who lives for another 12 years. The assumption is based
on conditional death probabilities from IBGE (for men and women) and their average age
difference, 9 years, in the survey. We select four profiles: skilled urban, unskilled urban, rural,
and government workers; for each of these types, we construct a predicted curve of real
earnings over a lifetime using PNAD cross-sectional data on median earnings by age.6 We
assume that workers who retire do not continue to earn wages from active work thereafter;
and that a postponement of retirement corresponds one-for-one to a lengthening of the
working life. In our examples, skilled workers are those who hold a university degree. Monthly
lifetime earnings in the profiles are expressed in gross terms, before taxes and social security
contributions.
10. The four types are illustrative, and by no means exhaustive. One could construct
other examples for women, for people who continue to work after retirement, for people who
have surviving children or no survivors at all, for people who move from the private to the
public sector halfway through their lives, for people who are caught in the transition between
the old and the new systems, and so on. The objective is not to span all the possibilities, and
not even to study “representative” cases in any statistical sense, but to develop a few examples
that can provide a sense of some of the distributional effects that can be expected of the
reform.
5
We are evaluating the most recent reform proposal (see Appendix for more detail on the changes in
parameters).
6
We estimate a Mincer regression with “Age” and “Age2” for each of our four types. There are significant
variations in wages in the survey, in particular for public workers.

BRAZIL
11. Overall gross lifetime income of the urban (private) sector skilled worker would
increase because of the reform. Without the reform, skilled workers with full contribution
history would retire on average at age 56 after 35 years of contribution. They would enjoy a
retirement benefit indexed to inflation till death and leave the full benefit to the spouse, who
would enjoy it for another 12 years on average. We consider now what would happen under
the proposed reform in the new steady state (that is, after the 20-year transition period is over)
to a similar individual. Because of the introduction of a minimum retirement age, the private,
skilled worker would postpone retirement. We assume that this individual would work for
45 years and retire at age 65. He would thus earn a wage for a longer time, enjoy a retirement
benefit for a shorter period, and leave a survivor’s pension to the spouse that would be lower
under the new rules.7 His net lifetime income, after accounting for taxes and contributions,
would also be higher than before the reform.
12. For workers with an incomplete contribution history, who already retire based on
old-age, the effect on the overall lifetime income from postponing retirement would be
negative but small. We assume that unskilled workers, some 65 percent of which earn up to
two minimum wages (Appendix I, Figure 1), contribute for 15 years currently but their
contribution time increases to 25 years after the reform. The period during which they receive
retirement benefits would remain the same but the benefit level would be lower because it
would be calculated over the entire contribution history (not the 80 percent highest wages, as
it is currently the case). Survivor’s pension would also decline.
Real Income of Urban Worker, Skilled Real Income of Urban Worker, Unskilled
(Constant R$ of 2015) (Constant R$ of 2015)
5,250 1,600
4,500 1,400
3,750 1,200
3,000 1,000
Before Reform 800
2,250 After Reform
600 Before Reform After Reform
1,500 Minimum Wage
400 Minimum Wage
750 200
0 0
20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90
Age Age
Sources: PNAD; and IMF staff calcualtions.
13. The reform would not affect the rural workers’ gross lifetime income. In the
survey, the median rural worker’s income is less than the minimum wage and about 45 percent
of rural workers earn half the minimum wage (Appendix I, Figure 1). The rural worker’s income
increases after retirement to one minimum wage. Since in the current reform the retirement
7
Under the new rule, survivors’ pensions are pro-rated to the size of the household; in the example, the sole
survivor would get a pension equal to 60 percent of the original benefit. Under the new rules the survivor’s
pension can be cumulated with the spouses’ own pension up to the amount of two minim wages. If this limit is
exceeded, the survivor can choose which one of the benefits to keep. In our example the surviving spouse has
no retirement benefit of her own, so this issue does not arise.

BRAZIL
age would not change for these workers, the reform would not impact their gross income.
Regarding their net income, the reform would have a small impact, to the extent they would
be required to contribute up to 5 percent of the minimum wage for at least 15 years.
14. Gross lifetime income of public sector workers would decrease. We assume that
the median public worker retires at age 60 under the current rules, in line with the average
observed for civil servants of the federal government—the group affected by the reform. After
reform, the public worker would continue to work for another 5 years. The retirement benefit
he receives would be lower, as in the case of private unskilled worker. The change in the net
lifetime income would be more pronounced, however, as the public worker would be
contributing to the social security system for a longer period (at an overall rate that is higher
than in the private sector).
15. From the perspective of the cumulative net transfers a person obtains from the
system, the proposed social security reform will reduce most significantly the differences
between public and private sector employees. The new rules governing minimum
retirement age and benefit calculations will not only generate fiscal savings, they will reduce
inequities in a social security system that currently benefit disproportionately a small share of
the population already enjoying higher wages—the public employees. The chart below shows
net cumulative transfers to each of the four types from the social security system, before and
after the reform. This is the difference between all retirement and survivor benefits on the one
hand, and all contributions paid by or on behalf of the employee, on the other hand. This
comparison naturally involved the calculation of present discounted values, using a 1 percent
real discount factor, in accordance with established OECD methodologies (Queisser and
Whitehouse, 2006). The calculation shows that the PDV of net transfers will be lower after the
reform for all four types of workers in the exercise; but the reduction will be greater for the
public-sector worker and the private sector skilled worker.
Real Income of Rural Worker Real Income of Public Worker

(Constant R$ of 2015) (Constant R$ of 2015)
900 3,500
800
3,000
700
600 2,500 Before Reform
500 Before Reform 2,000 After Reform
400 1,500
After Reform Minimum Wage
300
200 Minimum Wage 1,000
100 500
0 0
20 25 30 35 40 45 50 55 60 65 70 75 80 85 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90
Age
Age
Sources: PNAD; and IMF staff calcualtions.

BRAZIL
Change in the Lifetime Income Post Reform

(Constant R$ of 2015, Thousands)
300
200 Gross Net 1/
100
-100
-200
Urban Skilled Urban Unskilled Public Rural
Sources: PNAD; and IMF staff clculations.

1/ Earnings minus taxes and social security contributions paid by the worker.
PDV of Net Government Transfers

(Constant R$ of 2015, Thousands)
800
700 Before Reform
After Reform
600
500
400
300
200
100
0
Private Skilled Private Unskilled Public Rural

BRAZIL
Appendix I. Figure and Table

Figure 1. Brazil: Distribution of Wages Among Rural and Unskilled
Urban Workers 1/
(Percent)
Unskilled Urban Workers Rural Workers
35
50
30 45
40
25
35
Frequency
20 30
Frequency
25
15
20
10 15
5 10
5
0 0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Minimum wage Minimum wage
1/ To obtain the nominal wages the numbers in x-axis should be multiplied by the minimum wage.

46
BRAZIL
INTERNATIONAL MONETARY FUND
Table 1. Brazil: Pension Parameters Pre- and Post-Reform

Pension Reform Elements
Before the Reform After the Reform Transition

RGPS
Eligibility: (1) the sum of age and contribution time should be no
less than 85/95 for women/men; or (2) minimum of 30/35 years of
contribution for women/men.
Retirement by
Will be extinguished. There will only be retirement by age (see rules below).
Contribution Time Benefit value: in (1) is equal to s: the average of the worker's 80%
highest salaries; in(2) is equal to s*f , where f is the so-called
“fator previdenciário” (which depends on time of contribution, on
age and life expectancy at retirement). Eligibility for the transition rule: (1) minimum
age of 53/55 in 2017 for women/men. The
Eligibility: Minimum age of 62/65 years for women/men and minimum of 25 minimum age will gradually increase reaching
years of contribution time. 62/65 years for women/men in 2036/2038; and
Eligibility: minimum age of 60/65 years for women/men and (2) penalty linked to the time of contribution
minimum of 15 years of contribution time. Benefit value: is equal to s'*(70%x + 1.5%y1 + 2.0%y2 + 2.5%y3), where s' is still to (30% of the contribution time left to achieve
be determined (meanwhile it will be the average of all salaries since 1994), y1 is 30/35 years of contribution for women/men.)
Retirement by Age Benefit value: is equal to s*(70% + 1% per year of contribution), the number of contributing years within the 26 to 30 years range, y2 is the
where s is the average of the worker's 80% highest salaries. number of contributing years within the 31 to 35 years range, and y3 is the
number of contributing years after 35 years of contribution. Benefit value cannot
Contribution rate: employee = 8% to 11%; employer = 20%. be higher than s.
Contribution rate unchanged: employee = 8% to 11%; employer = 20%.
RPPS
Eligibility: Minimum age of 62/65 for women/men and 25 years of contribution
time. Minimum age for teachers will be 60 years old.
Eligibility: Minimum age of 55/60 years for women/men and Eligibility for the transition rule: (1) minimum
minimum of 30/35 years of contribution. Benefit value: is equal to s'*(70%x + 1.5%y1 + 2.0%y2 + 2.5%y3), where s' is still to age of 55/60 in 2017 for women/men. The
be determined (meanwhile it will be the average of all salaries since 1994), y1 is minimum age will gradually increase reaching
Benefit value: 80% of highest salaries. In 2013 the RPPS salaries the number of contributing years within the 26 to 30 years range, y2 is the 62/65 years for women/men in 2036/2038; and
Basic Retirement
were capped for new entrants to a value approximately equal to 6 number of contributing years inside the 31 to 35 years range, and y3 is the (2) pay a penalty in terms of time of
minimum wages. number of contributing years after 35 years of contribution time. Benefit value contribution: the penalty equals to 30% of the
Contribution rate: employee = 11%; employer (government) = cannot be higher than s. contribution time left to achieve 30/35 years of
22%. contribution.
Contribution rate unchanged: employee = 11%; employer (government) = 22%.
Table 1. Brazil: Pension Parameters Pre- and Post-Reform (Concluded)
BRAZIL
47
BRAZIL
References
Cuevas, A., Mulas-Granados, C., Karpowicz, I. and M. Soto, “Fiscal Challenges of Population
Aging in Brazil”, IMF Working Paper No. 17/99 (Washington: International Monetary Fund).
Queisser, M., and E. R. Whitehouse, 2006, “Neutral or Fair?: Actuarial Concepts and Pension-
System Design,” No. 40, OECD Publishing.
World Bank, 2017, Summary Note on Pension Reform in Brazil: “Why is it Needed and What
Will be its Impact?” World Bank Staff Note (Washington).

BRAZIL
INEQUALITY IN BRAZIL: A MICRO-DATA ANALYSIS1

In this study, we document the decline in inequality in Brazilian states in a new database
constructed from micro data from the national households’ survey. We find that labor
income growth, formalization, and schooling contributed to the decline in inequality
during 2004−14, but redistributive policies, such as Bolsa Família, have also played a
positive role. Going forward, it will be important to phase out untargeted subsidies, such
as public spending on tertiary education, and contain growth of public sector wages, to
improve budgetary efficiency and protect gains in equality.
A. Overview
1. Over the past decade or so, Brazil—a still highly unequal country—has been the poster
child for social mobility. According to the World Bank’s international poverty line, Brazil slashed
poverty from 25 percent of the population in 2004 to 8.5 percent in 2014. Extreme poverty declined
from 12 to 4 percent over the same period.2 As millions were lifted out of poverty, the middle class
was boosted. The commonly used inequality measure – the World Bank’s Gini coefficient (the closer
to 1, the more unequal) – declined from 0.60 in 1990 to 0.51 in 2014. Inequality reduction was
achieved thanks to a decade-long period of economic growth and deliberate income and social
inclusion policies, such as minimum wage increases and targeted social programs. Yet, inequality
remains high: based on data from the 2014 Pesquisa National de Amostra de Domicílios (PNAD),
labor income of the population in the top decile of the income distribution corresponds to
40 percent of labor income of all Brazilian families, the top 1 per cent receives about 12 percent, and
the top 0.1 per cent around 2.5 percent. Half percent of all labor income is concentrated in the top
0.01 percent.
2. The recession that started in 2014 is likely to have affected the pace of progress on the
social dimension. Earnings from work continue to be the main source of income for the poorest,
who are suffering disproportionately from job losses. Rising unemployment and compressed
households’ disposable incomes are affecting their living standards and jeopardizing social mobility.
Indeed, unemployment reached 13 percent in 2017, but has been higher for lower-skilled labor.3 But
even after the recession ends, the government will face a long period of fiscal consolidation. To
observe the cap on federal government non-interest expenditures, restraint will be necessary across
all categories of spending in the medium term.
3. In this paper, we construct a new database from disaggregated individual and

households survey data from the annual PNAD and use it to study the evolution and the
1
Prepared by Carlos Góes and Izabela Karpowicz (WHD).
2
Households whose income is less than $3.10 a day/less than $1.90 for extreme poverty at constant 2011 PPP-
adjusted U.S. dollars.
3
By early-2017, more than one in four young adults in Brazil were unemployed.

BRAZIL
drivers of inequality in Brazil. To our knowledge, we are the first to apply the methodological
techniques from the literature on global income inequality spearheaded by Milanović and his
co- authors, and recently updated in Lakner and Milanović (2015), to gain insights on both within-
and between-state inequality in Brazil. Another novelty of our paper is the use of a spatial price
differences index that we construct from PNAD data to allow comparability of nominal incomes
across states with unequal living standards. Because of the nature of the data employed, we focus
mainly on inequality of outcomes, and do not study inequality of opportunities, such as access to
education and health services, clear water and sanitation, and quality infrastructure.
4. We find that the decline in overall inequality in Brazil was driven by both falling
inequality within states and income convergence between states. We look at the evolution of
income and consumption patterns for specific income percentiles of the national income
distribution over time and show that income convergence was more evident around the median of
the distribution. From our regressions we find that most of the change in Gini can be explained by
income growth, higher schooling levels and labor formalization, but the targeted social program,
Bolsa Família, also contributed to income convergence. Civil servants’ wage growth has, in contrast,
slowed gains in equality. The reforms necessary to ensure fiscal sustainability should incorporate the
objective of improving spending efficiency while avoiding adverse effects on income distribution. As
labor formalization and income growth are slowing down, going forward, better targeting of social
benefits, rationalizing the tax system, and moderating civil servants’ wages will be key for preserving
gains in equality.
5. The paper is organized as follows: in Section B we describe the evolution of inequality in

Brazilian states and regions over the past decade using a novel data set; in Section C we present a
regression analysis to study the policy drivers of the decrease in inequality; and we conclude in
Section D.
B. Historical Trends in Regional Inequality 2004–14

6. In this section, we analyze the historical trends in inequality in Brazil based on the new
database constructed using micro-data from the Brazilian households’ survey (PNAD).
Because of the nature of the data we are using, we focus mainly on inequality outcomes, and do not
study inequality of opportunities, such as access to health, clear water and sanitation, and quality
infrastructure. We base our estimates of inequality on after-tax per capita income as reported in the
PNAD, which includes data on labor income and retirement benefits but excludes information on
other benefits, such as invalidity and survivors’ pensions, and income from financial and real assets.
7. Growth in incomes over the past decade has allowed the poorer segments of the
population to increase their consumption of durable goods. With access to electricity being
nearly universal across all income levels already in 2004, access to durable goods increased
substantially for all households over the following 10 years (Figure 1). But how have overall incomes
behaved and what is the state of income inequality today?

BRAZIL
Figure 1. Brazil: Convergence in the Consumption of Goods by Households

(In percent of total households in that quantile of the distribution)
Access to Electricity 2004 2014 Mobile Phones 2004 2014 Fridges 2004 2014
100 100 100
90 90 90
80 80 80
70 70 70
60 60 60
50 50 50
40 40 40
30 30 30
20 20 20
10 10 10
0 0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Percentile of national income distribution Percentile of national income distribution Percentile of national income distribution
Color TV 2004 2014 Washing Machines 2004 2014 PCs 2004 2014
100 100 100
90 90 90
80 80 80
70 70 70
60 60 60
50 50 50
40 40 40
30 30
30
20 20
20
10 10
10
0 0
0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
0 10 20 30 40 50 60 70 80 90 100
Percentile of national income distribution Percentile of national income distribution Percentile of national income distribution
Source: PNAD; and IMF staff calculations.
8. Income inequality in Brazil has declined. The Gini

Brazil: Gini Index
coefficient for Brazil published by the Brazilian Institute of (Index, 0= absolute equality)
0.56
Statistics (IBGE) fell from 0.54 in 2004 to 0.49 in 2014, and
0.55
other commonly used inequality measures also show 0.54
declining trends. We construct a Gini coefficient based on 0.53
the income reported by individuals and households in the 0.52

0.51
annual survey (PNAD) administered by the IBGE, adjusting
0.50
households incomes for spatial prices differences 0.49
throughout the country (Box 1). Our “adjusted” Gini index 0.48
has declined at the level of the country form 0.55 to 0.50 0.47
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
over the same period. The usefulness of this adjusted Sources: PNAD microdata; and IMF staff calculations.
income measure is in facilitating comparisons across states.
Box 1. The Cost of Living Adjustment

Inequality measures must take into account differences in the cost of living across and also within
countries to distinguish between nominal and real differences in incomes. Cross-country inequality
studies, such as Lakner and Milanović (2015) or Dollar and others (2013) for instance, typically correct the
between-country income statistics using PPP conversions, often based on national price indices. But
adjusting for living standards is important also when studying inequality within large countries because, as
highlighted by Deaton and Dupriez (2011), the Balassa-Samuelson effect may cause richer regions to show
permanently higher price levels. Indeed, price levels are not homogeneous in Brazil. Góes and Matheson
(2017) have documented large divergences of product-specific price dynamics, particularly for non-

BRAZIL
Box 1. The Cost of Living Adjust (Concluded)

tradables, across different metropolitan areas. Almeida and Azzoni (2016) showed that overall level price
differences in Brazilian metro areas can diverge with levels as low as -19 percent and as high as +14 percent
from the national average.
Because micro-data for consumer-price level differences is not available in Brazil we use information
on rental prices as a proxy. The consumer prices indices are available only for the 12 metropolitan areas,
insufficient to capture the potentially ample differences in living costs across Brazilian states. Using data on
declared households’ rent prices from the PNAD and other characteristics of the dwelling (such as the
number of rooms or area in square meters) we adjust households’ incomes for spatial price differences in a
two-step procedure. The advantage of using rental price data for the adjustment is that most of the price
dispersion generally comes from non-traded goods, and especially housing. Li and Gibson (2014), for
instance, have used data on dwelling sales in urban China to develop spatially-disaggregated indices of
house prices which they used as spatial deflators for both provinces and core urban districts.
First, for each sub-region 1,2, … ,7 ’ of each state 1,2, … ,27 ’ and each year 2004, … ,2015 ’,
we construct a rental spatial price difference index, which measures the percent deviation of the per room
average rental price from the national average:
, ,
, ,
, , ∗ 1
∗
where is the average monthly rent price for the cluster s,k, while is the average number of rooms per
household for the cluster, and the * denotes national averages.
Given that overall spatial price differences Brazilian Metro-Areas: Correlation Between Overall and Housing
Spatial Price Differences (In deviations from national averages)
can be well approximated by a linear
function of housing spatial price differences, 0.20
y = 0.5557x - 0.0118
Overall spatial price differences
we use the parameter from Azzoni and 0.15

R² = 0.7245
Almeida (2016), assumed to be homogenous 0.10
0.05
across regions, and our heterogeneous
0.00
rental spatial-price difference index to fit an
-0.05
overall spatial price difference index ̂ , ,
-0.10
, , Finally, we use ̂ , , to obtain adjusted
-0.15
households incomes, which are then used in
-0.20
the analysis of income distributions and their -0.25
trends. -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3
Housing spatial price differences
Our estimates of the overall Gini Sources: Almeida and Azzoni (2016); PNAD microdata; and IMF staff calculations.
coefficients are nearly perfectly Brazil: National Gini Coefficient (2004-14)
correlated with the official estimates of (Index, 0 = absolute equality)
the IBGE. Higher households income per 0.56

Authors' estimates
y = 0.9789x - 0.0108
capita regions tend to face price levels 0.55 R² = 0.9886
above the national average, while lower

0.54
income regions tend to face price levels
below the national average. Thus, adjusting 0.53
for spatial price differences compresses
0.52
nominal differences in incomes and
decreases the overall inequality indicator. 0.51
The estimated coefficient shown in the 0.5

figure (less than one) denotes the
0.49
compression effect of adjustment. 0.51 0.52 0.53 0.54 0.55 0.56 0.57 0.58
IBGE official estimates

BRAZIL
9. The decline in inequality was pronounced in the period studied, including among
regions. Within-state income distributions vary from state to state. In 2014, the Gini coefficient of
the most unequal state was 18 percent higher than the national Gini ratio, whereas the Gini of the
least unequal state was almost 20 percent lower than the national ratio. These differences are,
however, narrower than in the past: the standard deviation of state Gini coefficients decline from
0.035 to 0.033. Between-states inequality has decreased as a share of total inequality as incomes
grew faster in the poorer regions of the North, Northeast, and Midwest (blue, navy, and yellow lines
below). Convergence in average incomes led to a decreasing share of total inequality explained by
between-state inequality, as depicted also by the Generalized Entropy and Atkinson’s indices.4
Figure 2. Brazil: Income Inequality in Brazilian States: A Dynamic Decade

Brazil: Relative Gini Coefficient, by State, 2014 Brazil: Gini Coefficient, by State, 2004–2014
(State Gini / National Gini ratio) (State-wide Gini coefficient)
1.20 65
2014
1.15 2004
60
1.10
1.05 55
1.00
50
0.95
0.90
45
0.85
0.80 40
TO
DF
RJ
BA
RR
RS
AL
AP
SP
RN
SE
AC
SC
PI
PA
PB
RO
AM
PR
PE
CE
MG
MS
ES
GO
MA
MT
GO
RJ
RS
ES
AL
RN
AP
AC
RR
DF
RO
PE
PI
PA
AM
PR
PB
MG
MT
BA
MA
MS
CE
SE
SP
SC
TO
Sources: PNAD microdata; and IMF staff calculations. Sources: PNAD microdata; and IMF staff calculations.
Brazil: Real Income Per Capita Growth, by Region and Quantile, 2004–2014 Brazil: Trends in Between State Inequality
(Average real income growth per year; average across states per quantile; adjusted for spatial- (Share of total inequality represented by differences between state averages)
price differences)
8 5.5 18
GE(1) GE(2)
7 5 16
A(1), rhs A(2), rhs
6
4.5 14
5
4 12
4
North
3.5 10
3 Northeast
Southeast
2 3 8
South
1 Midwest
2.5 6
National
0
2 4
0 10 20 30 40 50 60 70 80 90 100
Percentile of state income distribution 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Sources: PNAD microdata; and IMF staff calculations.

10. Inequality within states also dropped. This was driven primarily by substantially higher
income growth rates for lower-income households in nearly all states. Inequality has declined
relatively more in the states with higher initial inequality in 2004, especially after excluding the
outliers (SC and DF), which illustrates convergence in within-state inequality indices across the
country.
4
The Generalized Entropy (GE) and Atkinson (A) indices, used as consistency checks, are perfectly decomposable into
within and between components.

BRAZIL
Figure 3. Brazil: Income Inequality in Brazilian States: Some Evidence of

Convergence
Brazil: Distribution of Within-State Gini Coefficients (2004-14)

(Index; 0 = perfect equality)
0.6
0.55
0.5
0.45
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Sources: Staff estimates with PNAD microdata. Box show interquartile range,
whiskers show 10th/90th percentiles.
Brazilian states: Evidence for Convergence in Within-State Income Inequality

(State Gini in 2004 and Percent Change in Gini Coefficient 2004–14)
5
Percent change in Gini coefficient (2004-14)
0 AP
y = -0.451x + 13.769 y = -0.991x + 43.497
R² = 0.1142 BA R² = 0.3026
-5 full sample SP RJ
DF
MS RR AC
MG
-10 SC RS RO AL
PA
MT SE
AM CE PB
-15 GO ES
RN
PR TO
PE no outliers
PI MA
-20
45 50 55 60 65
State Gini Coefficient in 2004
11. While convergence across states is most evident in the lower and upper quantiles of
the income distribution, some catch-up occurred also in the middle of the distribution. In the
figures below we explore how within-state income distributions fit into the national income
distribution. 5 Households belonging to the lowest and those belonging to the highest deciles of the
state income distribution also belong to the lowest/highest deciles of the national distribution. In
other words, the living standards of the lowest-earning and those of the highest-earning are similar
across states and regions. However, depending on the state, the state median household income
can fall anywhere between the 30th and the 70th percentile of the national distribution. These
differences have shrunk over time, as shown by a more pronounced downward shift of the curve
5
The different colors in Figure 4 represent households’ income per capita distributions of states that belong to the
region. For a legend of states see Appendix I.

BRAZIL
depicting the standard deviation of percentiles between the states and the national income
distribution around the 30th to 70th percentiles since 2004.
Figure 4. Brazil: Income Inequality in Brazilian States – Disaggregated
Brazil: Household Income per Capita Distribution, by State, 2004

(Percentiles of state-wide and nation-wide household income distribution, PPP adjusted)
100
Percentile of national income distribution
45 degree
90 Northeast
80 North
Southeast
70
South
60
Midwest
50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Percentile of state income distribution
Brazil: Household Income per Capita Distribution, by State, 2014

(Percentiles of state-wide and nation-wide household income distribution, PPP adjusted)
100
45 degree
90 Northeast
80 North
Southeast
70
South
60
Midwest
50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Percentile of state income distribution

BRAZIL
Figure 5. Brazil: Income Inequality in Brazilian States – Convergence in the Middle

Regional Inequality in Brazil, 2004–14 Brazil: Distribution of Households by Region and Income of the National
(Standard deviation of corresponding percentiles of national income distribution Distribution, 2014
per percentile of state-wide household income distribution, across states) (In percent of total households in each quantile)
12 100
national income distribution, across states
2014
2004 Midwest North Northeast South Southeast
10
Standard deviation of percentiles of
75
8
6
50
2 25
0
0 10 20 30 40 50 60 70 80 90 100 0
Percentile of state income distribution 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
Sources: PNAD microdata; and IMF staff calculations. Sources: PNAD microdata; and IMF staff calculations.
12. There is no evidence of reversal of progress with equality in the most recent PNAD
(2015). With the drop in employed population, real gross households earnings contracted in 2015
across all professions and for the first time in 11 years. The real income decline touched the entire
income distribution, but, as it was more severe in the higher income brackets, inequality fell slightly.
The official Gini index calculated for all income sources fell from 0.497 in 2014 to 0.491 in 2015. The
Gini calculated for labor income fell from 0.490 to 0.485 and, in the case of household income, from
0.494 to 0.493.
13. However, the continuation of the recession through 2016 may have dented equality
gains. Earnings from work represent a higher share of total income in the survey and a higher share
of the income of households in the lowest 25th percentile. As job destruction continued through
2016, and inflation remained high, the relatively poorer households have suffered more. Preliminary
2016 inequality estimates from FGV Social suggest that inequality widened slightly for the first time
in 22 years. The World Bank (2017) has estimated that the number of poor in Brazil will likely
increase by 2.5‒3.6 million by 2017, while the Gini index will increase from 0.51 to 0.52‒0.54. Among
the “new poor”, young, skilled workers in the service sector will represent the higher share of those
falling below the poverty line due to the crisis.
C. Macro-Policies and Inequality Outcomes: Regression Analysis

Rapid labor formalization has contributed to boosting living standards of the poor and so
have higher real wages and transfers. Redistributive taxation and social programs have
also played a role in the decline in inequality.
Tax policy
14. If achieved through progressive taxation, increases in tax revenues can be correlated
with declining inequality. Brazil’s overall tax system relies relatively more on indirect taxes, which

BRAZIL
are regressive.6 Effective personal income tax (PIT) rates, which take into account all the admissible
deductions (green line in the chart), do not seem progressive either. However, taking into account
the taxation of dividends at corporate level the system’s progressivity appears to be restored (red
line). Applying standard benefit-tax incidence analysis, Lustig and others (2014) find that personal
income taxes in Brazil are progressive and redistributive, and contributed to reducing the Gini of
after-tax incomes by 1.9 percent in 2009.
Minimum wage and expenditure policies
15. Supported by strong growth, the minimum wage policy in Brazil has sustained upward
social mobility for lower classes. The effect of the minimum wage policy on inequality is
ambiguous, given its potentially offsetting impact on employment and inflation (Jaumotte and
Osorio Buitron, 2015). According to Maurizio (2014), increases in the minimum wage in Brazil led to
wage compression, which helped to reduce inequality among wage earners. Indeed, the hourly
wage for a worker with a given level of education rose much faster among the poor than for the rest
of the population in the last decade mainly because of the minimum wage policy. This occurred as
the increase in the real minimum wage above productivity was accompanied by a decline in the
unemployment rate supported by Brazil’s relatively strong output growth during the period.
However, in an environment of low or negative growth, such as the one facing Brazil recently, the
minimum wage growth may be affecting employment negatively, with the effect being more
pronounced for the unskilled workers (IMF, 2015). This would in turn lead to higher before tax (or
gross) inequality.
16. Public sector wage increases systematically above private sector wage growth appear
to have slowed equality gains. While it is not uncommon to find evidence of public sector wage
markups in the literature (Clements and others, Brazil: Change in Real Annual Earnings, per Sector
2010), the wage premia on public sector jobs in (Accumulated change in real earnings, per sector, in percent)
60
Brazil are particularly high. In 2014, the estimated Private Sector
50
median premium on public sector jobs across Public Sector
40
comparable professions was about 50 percent up
30
to the secondary education level (Box 2). To the
20
extent that public sector workers’ incomes are
10
higher that private sector workers’, stronger
0
growth of wages in the public sector may have
-10
moderated equality gains achieved in the recent
Jan-04
Sep-04
May-05
Jan-06
Sep-06
May-07
Jan-08
Sep-08
May-09
Jan-10
Sep-10
May-11
Jan-12
Sep-12
May-13
Jan-14
Sep-14
May-15
Jan-16
decade.
Sources: PME/IBGE.
6
The ratio of direct to indirect taxes at the general government level was 45 percent in 2016. Brazil relies on indirect
taxes more than other Latin American economies and significantly more than OECD countries (OECD, 2010).
According to Amaral and others (2016), the average Brazilian worker pays 15 percent of his gross income in income
taxes, 3 percent in asset taxes, and 24 percent in consumption taxes. Those making up to R$3,000 per month pay
24 percent of their gross income in consumption taxes, while those making more than 10,000 pay 17 percent in
consumption taxes. While personal income taxes are progressive, excessive reliance on consumption taxes makes the
overall system regressive.

BRAZIL
17. Bolsa Família is Brazil’s flagship Brazil: Fiscal Cost and Reach of Bolsa Família
(Percent of GDP and millions of households)
social assistance program for reducing 0.6
<— Begin of
poverty. Beneficiary coverage has increased
Fiscal Cost
Brasil Sem Miseria
(In percent of GDP)
from about 6.5 million households in 2004,

2014
2015
2013
when it was founded, to over 14 million in

2016
0.4 2012
2011
2014 (56 million people). Budgetary
2007
2009 2010
2008
appropriations for the program have also
2006
2005
increased from about 0.3 percent of GDP to 0.2
2004
0.6 percent of GDP over the same period.
The World Bank (2017) estimates that Households
(millions)
0.0
58 percent of the decline in extreme poverty 6 7 8 9 10 11 12 13 14 15
in Brazil over 2004‒14 was due to Bolsa Sources: Ministry of Social Development; and IMF staff calculations.
Família transfers. Soares and other (2006)
report that in 2005 about 80 percent of the Bolsa Família budget went to families below the poverty
line (half of the minimum wage per capita) and that the program was responsible for 21 percent of
the decline in the Gini coefficient between 1995 and 2005.
Box 2. Returns to Education and Public-Private Wage Gap

We estimate the returns on education in Brazil by means of two “Mincer” regressions (Mincer, 1974) with
identical specifications that relate the log of wages to years of schooling and experience for the public and
the private sector separately, for each period 2004, … ,2015 ′ . The model contains more than 50 other
controls:
ln , , , , , , , ∑ , , , ∑ , , , ∑ , , , ∑ , , ,
∑ , , , ,
where , is the monthly wage for person i at period t; , is the years of formal schooling; , is the years
of experience (defined as the individual’s age minus years of schooling minus 6—the age when mandatory
education starts); , is a gender dummy; , , are dummies for races; , , are dummies for occupations
(formal/informal worker, military, civil servant, domestic worker, self-employed, etc.); , , are dummies for
sectoral economic activities (agriculture, industry, manufacturing, construction, commerce, etc.); , , are
dummies for worker’s class (director, middle management, administrative, sales, etc.); , , are dummies for
the Brazilian states; and , is the error term.
In the second step, we use the estimated coefficients from the regressions to generate two vectors of fitted
values for each one of the ~150 thousand workers in the sample who belong to either the private or the
public sector. The fitted values show the expected log wages for individuals, given the same set of
observable characteristics.

BRAZIL
Box 2. Returns to Education and Public-Private Wage Gap (Concluded)

We find that predicted earnings are an Brazil: Predicted Returns on Schooling, 2014
(In log of monthly wages given years of formal education; distribution of predicted wages
increasing function of the years of schooling in given 50+ controls for geography, demographics, experience, education, industry,
Brazil for public as well as private sector jobs, management level, etc.)
9
but earnings among those in public sector jobs 8.5
are consistently higher, in line with Souza and 8
Medeiros (2012). Up to the secondary 7.5 Public Sector
education level, the 25 percent lowest predicted 7
earnings in the public sector are higher that the 6.5

6 Private Sector
median earnings in the private sector.
5.5
We define the public sector wage premium as
Public
Public
Public
Public
Public
Public
Public
Public
Public
Public
Public
Public
Public
Public
Public
Public
Private
Private
Private
Private
Private
Private
Private
Private
Private
Private
Private
Private
Private
Private
Private
Private
the difference between the two predicted 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
values , ln , ln , . Like
1
Sources: PNAD microdata; and IMF staff calculations. Boxes show interquartile range. Whiskers show
10th/90th percentiles. We ran Mincerian regressions with the same controls for a sample of ~150,000 obs,
Braga and others (2009), we find some signs of calculated predicted values from coefficients resulting from the private and public sectors, and ploted the
distributions.
compression of the premium at higher
educational levels. However, we find the premium Brazil: Civil Servants Median Wage Premia, 2014
(In percent difference over predicted private sector wages)
to be high across all years of schooling, which was
not the case ten years ago, when they published 60
their study. 55
50
Given their observable characteristics, at least 75 45
percent of workers would benefit today by moving 40
99%
confidence interval
from the private to the public sector in comparable 35
jobs at all education levels. 30
_________________________________ 25
1 20
We ran Mincer regressions with the same controls for a
sample of ~150,000 observations, calculated predicted 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Years of schooling
values from coefficients resulting from the private and
public sectors. Standard errors are from a bootstrap
simulation with 500 repetitions.
18. Even when they don’t worsen income inequality by themselves, implicit education
subsidies use up resources that could be otherwise employed to improve income equality.
While spending on education may be concomitantly pro-growth and pro-equality (Ostry and others,
2014), it can be also seen as a blanket subsidy and weakly targeted transfers generally constitute an
inefficient use of scarce resources. In Brazil, public universities are tuition-free. They are also more
accessible to children of wealthier parents, who often have studied in private primary and secondary
schools (World Bank, 2016). Indeed, nearly half of the public university student population in the
PNAD survey belongs to households in the top quartile of the income distribution, while only
9 percent of university students come from families in the bottom quartile (see Box 3). This type of
implicit subsidy benefits the rich disproportionally. Equality of opportunities could be enhanced by
redistributing resources away from tertiary education towards improving provision of early
childhood and primary schooling which would improve overall spending progressivity (IMF, 2014).

BRAZIL
Box 3. An Example of Poorly Targeted Transfers: Public Universities

Over the past decades, the educational level of Brazil: Share of Population With Secondary Education
(Distribution of Brazilian municipalities according to share of population with at least
Brazilians improved significantly. The share of secondary education)
population between 20 and 22 years old that 25
Share of municipalities in range

completed at least secondary education increased
20
from 45.5 percent in 2004 to 60.8 percent in 2014,
1990 2000 2010
according to data from the PNAD. The improvement 15
was widespread across regions, as depicted by the

10
curves shifting to the right in the figure. At the same
time, Brazil has expanded access to tuition—free, tax 5
payer-funded public universities. Between 2000 and
0
2014 the number of students in public universities
0 20 40 60 80
more than doubled—from 0.89 million to Share of 25+ years-old population with at least secondary education
1.96 million (Ministry of Education, INEP, 2015). Sources: IMF staff calculations with census data compiled by UNDP.
Students from better-off households are Brazil: Overrepresentation of Top Quartile in Public Universities, 2014
(Share of overall population and public unversity population which belong to each
overrepresented in public universities. Nearly quartile of household income per capita)
half of the public university student population in 100%
Bottom Quartile
the sample belongs to households in the top 90% 25%
80%
quartile of the income distribution, while only 47%
70%
9 percent of university students comes from the 60% 25%
Second Quartile
bottom quartile. Meanwhile, about 40 percent of 50%
the younger cohorts of the Brazilian population still 40% 25%

25%
Third Quartile
30%
fails to complete secondary education. 20% 19%
25%
We provide a more robust proof of the relationship 10%
9%
Top Quartile
0%
between income and access to public universities Overall Population Public University
by as in Góes and Duque (2016). We estimate a Population
logit model with PNAD data to obtain the
probability of being public university student for individuals between 17 and 24 years old conditional on
household’s income per capita and a set of controls:
Pr 1| , , ∈ 0,1
1
where is a categorical variable denoting a student Brazil: Probability of Attending Public University and Household 1/
currently in a public university for individual , is Income, 2014 (In percent, by household income per capita)
household family income per capita and is a set of 60
Probability of attending public university
individual controls—which include age, gender, race 50

and regional dummies (see Appendix I, Table 3).
40
We find that, even after controlling for 30

geographic and demographic characteristics,
20
students from richer households are considerably
more likely to attend public universities. In fact, a 10
student in the 25th percentile of the income 0

distribution has a 2 percent probability of attending a 0 5,000 10,000 15,000 20,000 25,000 30,000
Household income per capita, in Brazilian reais per month
public university while the one in the 99 percentile
th
has more than 30 percent probability of attending it. 1/ Average predicted values from multivariate logit regression for different income cohorts.
This finding is consistent with the intuitive diagnostic that children with richer parents, who can afford to
study in private primary and secondary schools, obtain easier access to publicly-funded universities (World
Bank, 2016).

BRAZIL
Box 3. An Example of Poorly Targeted Transfers: Public Universities (Concluded)

Redirecting government spending from tertiary to primary and secondary education would
improve overall welfare and equality. Funding a student at the higher education level costs
about four times as much as funding a student at the secondary education level in Brazil. This
ratio is much higher than the OECD average of 150 percent (OECD, 2014). Given that many
Brazilians do not complete secondary education and the rate of return to investment in human
capital tends to be higher at lower levels of education (Heckman, 2008), targeting education
spending on the poor and cutting subsidies to the rich can generate fiscal savings while making
the access to education fairer, and ultimately equality of opportunities better.
19. Below we present the results from the analysis of the drivers of changes in inequality
in a regression framework. We use our data set constructed from the PNAD by aggregating
individual data into state-level statistics.7 We complement the state-level data with information on
annual Bolsa Família state budgets and federal income tax revenues collected in states.8 We use a
two-stage least squares approach with state dummies to control for time-invariant state-specific
characteristics. First, we regress the income of the top and bottom quartiles on the average income,
civil servants’ income, tax revenues, the share of formal sector workers in total, employment,
schooling, and the per capita Bolsa Família budget for each state. We have two regressions in the
first stage because the independent variables of interest can have different effects on different
points of the income distribution—e.g., the average income growth rate can have a stronger impact
on the bottom quartile than the top quartile growth rates. We then regress the Gini coefficient on
the predicted values from the two first stage regressions. This allows us to pin down both the
marginal impact (through the different quartiles) as well as the net impact of the first stage variables
on the Gini coefficient (see Appendix II for more detail).9
Results from Stage II regression

Brazil: Decomposition of HH Income Inequality Dynamics (2004–14)
(In Gini Points)
0.10
0.05
0.00
-0.07
-0.06 -0.06
-0.05
-0.05 -0.04
-0.10 -0.06
-0.15
Brazil North Northeast Southeast South Midwest
Income, bottom quartile Income, top quartile Total
Source: PNAD; and IMF staff calculations.
7
See Appendix 2 for more details on the data sources and the construction of variables.
8
We use the per capita income for the household as reported in the survey which includes data on labor income and
retirement benefits but excludes information on other benefits, such as invalidity and survivors’ pensions, and income
from financial assets.
9
Following this procedure takes care of potential endogeneity problem in the second stage regression.

BRAZIL
20. We find that employment, labor formalization, income growth, Bolsa Família budgets
and schooling have contributed to declining inequality. The variables in the regression (Table 1)
are mostly significant and bear the expected sign. Together, increased schooling and labor
formalization explain the largest share of the decline in the Gini, but growth of average incomes and
Bolsa Família also contributed to lowering inequality.10 In contrast, the growth of incomes of civil
servants has affected equality negatively. Income taxes are not a significant determinant of
inequality, possibly because the PNAD may be underestimating the income of the top 1 percent of
the population (see below).
Table 1. Brazil: Coefficients from the 2SLS Procedure

Coefficients from the 2nd Stage Regression
LN of average HH income per capita, bottom quartile (1) -0.186 *** (1)
LN of average HH income per capita, top quartile (2) 0.198 *** (2)
constant -0.057
Marginal Effects on Gini from the 1st stage specification 1/ (1) (2)
LN of average HH income per capita -0.014 *** ***
LN of Civil Servant HH income per capita 0.054 *** ***
Share of Civil Servants, pct. -0.001
LN of Bolsa Familia exp. per capita -0.007 *** ***
Share of Formal Workers, pct. -0.002 ***
Employment Rate, pct. -0.001 ***
Avg Schooling Years for Adults, bottom quartile -0.030 ***
Avg Schooling Years for Adults, top quartile 0.004
Tax Revenues, pct 0.003
1/ The asterisks denote significance at the first stage in relation to first stage dependent
variables (1) or (2). The coefficients shown here are the combination of 1st and 2nd stage
regression coefficients.
Sources: PNAD; Receita Federal; Ministério de Desenvolvimento Social; IMF staff calculations.
21. Potential data limitations must be kept in mind when interpreting results from our
study. It is important to note that in the PNAD income received by the population in the top
1 percent of the income distribution in Brazil may be underreported (Souza, 2013). If income of the
richest segment of the population not captured in the PNAD has increased over time, we could
overestimate the decline in inequality in the country. In a recent study by Medeiros and others
(2015), the authors combine data from the personal income tax returns (DIRPF) in Brazil with the
PNAD and conclude that a growing share of income was received by the top 1 percent between
2006 and 2012 which has caused overall inequality to stagnate over that period. However, such
adjustment was not possible in our study as the DIRPF data is not available for the entire period
under consideration. Implicit in our approach is also the assumption that estimated parameters are
homogeneous and linear across states. This is a limitation by construction which implies that
extrapolating results from our estimates to draw conclusions for the future or for specific states
must be done cautiously.
10
The findings on Bolsa Família are in line with previous literature in Brazil which underlines the redistributive power
of the program (Neri, 2010; Azzoni and Silveira-Neto, 2012).

BRAZIL
D. Conclusions
22. In this study, we find evidence of a decline in inequality in Brazil, both between as well
as within its 27 states from 2004 to 2014, and document the drivers of this phenomenon.
Falling inequality can be attributed to convergence in households’ incomes in the proximity of the
mid-point of the state-wide distribution, which was stronger in more unequal states. We find that
growth in incomes and labor formalization all contributed to declining inequality, but faster wage
growth in the public sector played the opposite role. In terms of redistribution policies, Bolsa Família
budgets were progressive, and so were higher schooling levels among the poor.
23. Against a backdrop of a recession that is eroding incomes of the poor, the policy
framework will need to strike a balance between the goals of fiscal sustainability and income
equality. The sharp recession has brought to the fore important reform priorities, such as social
security, labor market and tax reforms, some of which are already at an advanced stage. Preserving
equality gains and moving forward with the inclusiveness agenda will remain key for gathering
support for these reforms. This can be achieved without further increases in spending, by
moderating civil servants’ wage growth and using direct instruments to provide benefits based on
need (IMF, 2014), such as Bolsa Família. Improved access to education and educational attainment
for lower-income families can be achieved by redirecting resources currently funding universal,
tuition-free access to tertiary education, while continuing to support university students based on
need. The tax system can rely more on direct taxation and less on indirect taxes. Finally, the
minimum wage policy should provide appropriate remuneration for the poor without discouraging
formal employment.

BRAZIL
Appendix I. Data and Sources

Legend – States
Decline in Inequality in States 2004−15 Region Acronym State

AC Acre
AM Amazonas
AP Amapá
RR North PA Paraná
AP
RO Rondônia
RR Roraima
TO Tocantins
AM PA
MA CE AL Alagoas
RN
BA Bahia
PI PB
PE CE Ceará
AC
TO
AL
MA Maranhão
RO SE
Northeast PB Paraíba
BA
MT PE Pernambuco
PI Piauí
GO DF
RN Rio Grande do Norte
MG SE Sergipe
ES
MS ES Espírito Santo
SP RJ MG Minas Gerais
Southeast
North RJ Rio de Janeiro
PR
Northeast SP São Paulo
Midwest
SC Southeast PR Paraná
South
South SC Santa Catarina
RS
RS Rio Grande do Sul
MS Mato Grosso do Sul
MT Mato Grosso
Midwest
GO Goiás
DF Distrito Federal
Gini (2004) Gini (2014)

0.60 0.60
0.55 0.55
0.50 0.50
0.45 0.45
Sources: PNAD, and IMF staff calculations.

BRAZIL
A. Data Sources and Definition of Variables used in the Regression

The Households Survey—PNAD
Most of our data consists of a novel dataset constructed by the authors from PNAD microdata.
PNAD (Pesquisa Nacional por Amostra de Domicílios) is a National Household Survey conducted on
a yearly basis. It collects data on nearly 360,000 individuals distributed through about 140,000
households. In total, to build time series for each state between 2004−15, we used around 6 million
data points.
The PNAD has two annual datasets. The first one presents collective characteristics of each
household while the second one has specific characteristics of each individual. We incorporated the
complex survey design of PNAD by using the weights that the survey provides for the relative
representativeness of each household/individual and adjusting our estimates and reported errors by
double clustering at the state and household levels.
Using the household database, we constructed household income per capita quantiles for states and
assessed the dynamics of household characteristics, including consumption patterns, for each
quantile. We also constructed time series, for each state, of inequality indices (Gini, General Entropy
and Atkinson), based on household income per capita.
Using the individual database, we constructed time series, for each state, for the following
indicators: average household income per capita; average wage per active worker; share of formal
workers in total population; average years of schooling for adult population (above 16); share of
population between 17 and 30 years old in private universities; share of population between 17 and
30 years old in public universities, per capita Bolsa Família budget.
Definition of Variables used in the Baseline Regression
 Data on Bolsa Família budgets by state comes from the Brazilian Ministry of Social
Development. We used these data to calculate the yearly real per capita outlays. The variable
Bolsa Família was constructed by dividing the total Bolsa Família state budget by the state
population after adjusting for spatial price differences and inflation. The indicator adjusted for
spatial price differences.
 The tax revenue variable is constructed the federal personal and corporate income tax revenues
collected by the states and reported by Receita Federal, divided by the state GDP from IBGE.
 Approximately between 40 and 80 percent of workers are “informal” in the dataset, depending
on the state in which they reside. The formal work variable was constructed from the answers in
the PNAD and expressed as the share of formal workers in total for each state:

BRAZIL
 “formal worker”
- Formal contract (carteira assinada);
- Military;
- Civil servant;
- Employer/entrepreneur;
- Domestic employee with a formal contract (doméstico com carteira assinada);
- Unpaid/Voluntary work (não remunerado);
- Self-employed as a manager/director;
- Self-employed as an artist.
 “informal worker”
- No formal contract (sem carteira assinada);
- Domestic with no formal contract (doméstico sem carteira assinada);
- Self-employed (except for manager/director and artist);
- Self-consumption worker in production;
- Self-consumption worker in construction.
 The employment rate variable is expressed as the share of persons who were working in the
reference week in the total labor force (persons economically active, i.e. employed or actively
looking for job).
 The average income variable is the average household per capita income for the state
aggregated from the household income indicator available in the survey and adjusted for spatial
price differences.
 The civil servant income is equal to the household income per capita for households whose
reference person (respondent) is a civil servant. The indicator is adjusted for spatial price
differences.
 The schooling variables is the average number of years of schooling for persons in the lower
and top quartiles of the national household income per capita distribution.

BRAZIL
Figure 1. Brazil: Household Income Per Capita Distribution, 2014

(Percentiles of state and national income distributions, adjusted for spatial price differences)
North Northeast
100 100
RO

MA
90 AC 90 AL
80 AP PI
RR 80
70 CE
PA 70
RN
60 AM
60 PB
TO
50 45 degree 50 SE
40 BA
40
45 degree
30 30
20 20
10 10
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Percentile of state income distribution Percentile of state income distribution
Southeast South
100 100

MG
90 PR
ES 90
SC
80 RJ 80 RS
70 SP
70 45 degree
45 degree
60 60
50 50
40 40
30 30
20 20
10 10
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Midwest Alagoas vs. Distrito Federal

100 100
MS DF
90 90
MT
80 80 AL
GO
70 70 45 degree
DF
60 45 degree 60
50 50
40 40
30 30
20 20
10 10
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Source: PNAD microdata; and IMF staff calculations.

BRAZIL
68
Table 1. Brazil: Percentile of National Per Capita Income Distribution for Percentile of State-Wide Per Capita
Income Distribution, 2014

Percentile of state-
wide income per 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
capita distribution
RO 5.0 9.4 13.5 18.3 22.0 26.3 31.9 35.7 42.2 46.3 50.6 56.9 61.3 67.5 73.1 79.4 83.6 88.5 92.7 98.5
Appendix II. Econometric Results and Tables

AC 5.0 5.0 8.7 10.0 13.1 15.8 21.3 25.0 29.2 34.5 39.7 47.4 53.2 57.4 63.6 71.5 79.0 86.3 93.2 98.9
AP 5.0 5.0 6.5 10.0 11.7 15.0 19.1 21.4 26.5 30.0 34.8 40.0 45.0 50.2 58.0 66.3 74.7 82.9 91.4 98.2
RR 5.0 9.9 12.4 17.2 20.3 25.0 29.8 34.3 38.3 43.8 49.6 51.9 56.7 65.3 71.2 78.8 86.6 91.0 95.0 99.7
PA 5.0 5.0 9.4 10.0 14.4 17.0 20.0 24.8 28.4 32.2 36.4 41.9 48.8 55.0 58.4 64.2 72.1 80.5 88.2 97.3
AM 5.0 10.0 12.1 16.3 20.0 23.9 28.0 30.9 36.3 43.9 50.0 54.8 61.8 70.0 76.1 82.4 87.3 90.0 94.3 98.1
TO 5.0 9.7 14.6 19.9 24.8 28.0 32.6 37.1 44.0 53.8 n.a 57.8 61.9 67.3 73.2 77.7 83.9 89.1 93.6 98.8
MA 5.0 5.2 10.0 11.8 15.0 19.7 23.4 25.8 30.7 35.0 40.6 47.3 54.9 60.0 62.7 68.6 75.3 82.6 90.6 98.1
PI 5.0 7.1 11.6 16.6 21.2 26.2 30.0 34.2 41.4 45.5 51.2 58.0 64.5 66.1 71.2 77.7 80.5 87.3 92.8 98.8
CE 5.0 5.0 9.6 13.1 16.4 20.2 25.0 29.0 32.6 36.7 42.5 49.6 58.7 60.0 62.9 71.0 75.0 81.3 88.5 97.5
RN 5.0 7.9 12.3 16.9 22.2 25.0 30.0 33.2 38.9 44.0 49.5 56.5 60.0 63.0 68.5 76.5 81.0 86.8 93.1 99.3
PB 5.0 8.6 13.3 17.1 21.1 25.0 30.0 33.7 39.4 44.7 51.8 59.5 n.a 63.4 69.3 77.4 83.2 88.9 94.3 99.9
PE 5.0 6.1 10.0 12.0 15.6 20.0 24.4 28.3 32.2 37.0 42.6 49.3 51.8 56.7 62.9 70.3 75.4 82.1 89.5 98.0
AL 5.0 5.0 8.6 10.9 15.0 18.4 21.6 25.0 30.0 33.0 38.2 44.3 52.7 59.9 61.8 66.8 73.8 80.6 88.5 97.4
SE 5.0 9.6 12.6 15.0 19.6 24.7 29.0 32.1 36.9 42.3 47.5 53.8 59.6 61.8 67.6 71.8 77.6 83.9 90.0 97.7
BA 5.0 5.7 10.0 13.6 16.7 22.2 25.0 29.7 33.6 38.2 44.0 53.6 55.0 59.1 65.3 71.5 77.8 84.9 91.8 98.7
MG 6.4 12.4 18.8 23.8 29.4 34.9 40.9 48.7 50.0 53.5 58.3 62.7 68.1 71.8 76.7 81.2 85.7 90.1 95.1 100.0
ES 7.1 13.0 19.2 25.4 31.3 35.6 41.7 49.0 51.7 56.7 62.1 67.7 71.4 76.3 81.2 85.0 88.8 92.2 96.0 100.0
RJ 6.0 10.6 15.0 19.5 24.1 29.2 34.3 35.4 41.2 45.0 50.9 56.7 62.9 67.7 73.2 78.3 85.0 91.2 95.9 100.0
SP 7.2 13.5 18.3 24.7 29.6 35.1 40.0 42.1 46.9 53.5 58.9 63.6 68.0 73.2 78.3 82.9 87.4 91.7 95.9 100.0
PR 8.4 17.4 23.5 30.4 36.6 43.0 45.4 50.2 55.5 60.2 65.0 68.7 73.1 77.2 80.0 84.8 88.5 92.1 96.0 100.0
SC 12.4 24.0 32.8 40.3 45.0 48.8 55.0 59.7 64.5 67.9 70.1 75.0 78.3 80.0 84.7 87.4 90.0 93.1 95.9 100.0
RS 7.3 15.8 22.5 29.5 36.7 44.2 45.0 51.4 56.2 61.7 65.0 70.0 74.6 79.3 82.2 85.4 90.0 93.2 96.6 100.0
MS 8.1 16.5 22.4 27.4 32.0 37.9 43.2 49.2 50.5 55.8 60.2 65.2 70.3 75.0 79.4 83.6 87.9 92.2 95.9 100.0
MT 8.4 17.4 23.7 29.0 34.0 40.4 45.0 49.3 52.8 57.9 63.6 67.0 71.3 75.0 80.0 84.0 87.6 91.6 95.0 99.6
GO 7.9 14.1 20.8 27.1 32.5 37.8 43.3 49.6 n.a 55.3 60.3 65.2 70.0 73.7 77.9 82.2 85.5 90.0 94.4 99.1
DF 6.3 11.7 17.4 21.9 28.9 34.3 38.2 46.1 54.0 62.1 67.9 74.2 82.0 85.5 90.0 94.9 96.4 100.0 100.0 100.0
Range (Max - Min) 2 5 6 8 9 9 10 10 10 11 10 10 9 9 8 7 6 5 3 1

Table 2. Brazil: Drivers of Income Inequality in Brazil’s States from the 2SLS Regression (2004–14)
1st Stage Least Squares Results

LN of average HH income per capita, bottom quartile (1) Coeff. Std. Err z P>|z| [95% Conf Interval
LN of average HH income per capita 0.734 0.061 12.040 0.000 0.614 0.853
LN of Civil Servant HH income per capita -0.135 0.040 -3.380 0.001 -0.213 -0.057
Share of Civil Servants, pct. -0.002 0.004 -0.570 0.569 -0.010 0.005
LN of Bolsa Familia exp. per capita 0.134 0.021 6.480 0.000 0.093 0.174
Share of Formal Workers, pct. 0.002 0.002 1.210 0.227 -0.001 0.006
Employment Rate, pct. 0.005 0.003 2.060 0.039 0.000 0.011
Avg Schooling Years for Adults, top quartile 0.156 0.019 8.280 0.000 0.119 0.193
Avg Schooling Years for Adults, bottom quartile -0.007 0.011 -0.620 0.533 -0.028 0.015
Tax Revenues, pct -0.013 0.009 -1.450 0.146 -0.031 0.005
LN of average HH income per capita, top quartile (2) Coeff. Std. Err z P>|z| [95% Conf Interval
LN of average HH income per capita 0.618 0.056 11.130 0.000 0.510 0.727
LN of Civil Servant HH income per capita 0.145 0.036 3.990 0.000 0.074 0.216
Share of Civil Servants, pct. -0.005 0.004 -1.350 0.176 -0.012 0.002
LN of Bolsa Familia exp. per capita 0.090 0.019 4.780 0.000 0.053 0.127
Share of Formal Workers, pct. -0.010 0.002 -5.890 0.000 -0.013 -0.006
Employment Rate, pct. 0.001 0.002 0.210 0.832 -0.004 0.005
Avg Schooling Years for Adults, top quartile -0.003 0.017 -0.200 0.842 -0.037 0.030
Avg Schooling Years for Adults, bottom quartile 0.015 0.010 1.490 0.137 -0.005 0.034
Tax Revenues, pct 0.003 0.008 0.380 0.707 -0.013 0.019
Sources: PNAD; Secretaria da Receita Federal; Previdência Social; and IMF staff calculations.
BRAZIL
69
BRAZIL
Table 3. Brazil: Logistic Regression: Determinants of Attending Public University, 2014 1/

Marginal Effects on Probability of Attending Public University
(1) (2)
0.025*** 0.022***
LN of HH income per capita
(0.001) (0.001)
0.001***
Age
(0.000)
-0.004***
Male
(0.001)
-0.014***
Black
(0.003)
-0.013***
Brown
(0.002)
0.016**
Asian
(0.007)
-0.024*
Native Brazilian
(0.013)
0.021***
North Dummy
(0.002)
0.021***
Northeast Dummy
(0.001)
-0.000
South Dummy
(0.002)
0.013***
Midwest Dummy
(0.002)
Pseudo-R2 0.0716 0.0953

Observations 40,385 40,385
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
1/ Coefficients denote the expected marginal response in the dependent variable after a unit change in the independent variable at specific points.
Continuous variables (age and income) have been calibrated such that coefficients denote the marginal responses at their respective medians.
Categorical variables (gender, race, and state dummies) have been calibrated to zero and the coefficients denote the marginal responses after a
change from zero to one in the categorical variable. Households’ income per capita was adjusted for spatial price differences.

BRAZIL
References
Almeida, A. N., and C.R. Azzoni, 2016, “Custo de vida comparativo das regiões metropolitanas
brasileiras: 1996‒2014,” Estudos Econômicos, São Paulo,Vol. 46(1), pp. 253‒76.
Azevedo, J. P, David, A.C., Rodrigues Bastos, F. and E. Pineda, 2014, “Fiscal Adjustment and
Income Inequality: Sub-National Evidence from Brazil,” IMF Working Paper No. 14/85,
(Washington: International Monetary Fund).
Azevedo, J.P, Inchauste, G. and V. Sanfelice, 2013, “Decomposing the Recent Inequality Decline
in Latin America, The World Bank, Europe and Central Asia Region, Poverty Reduction and
Economic Management,” Working Paper No. 67/15, December.
Braga, B., Firpo, S. and G. Gonzaga, 2009, “Escolaridade e o diferencial de rendimentos entre o
setor privado e o setor público no Brasil, pesquisa e planejamento econômico (PPE),” Vol. 39,
No. 3, pp. 431‒64.
Clements, B., Gupta S., Karpowicz, I. and S. Tareq, 2010, “Evaluating Government Employment
and Wages,” IMF Technical Notes and Manuals 2010/15, (Washington: International Monetary
Fund).
Deaton, A. and O. Dupriez, 2011, "Spatial Price Differences Within Large Countries,"
Manuscript, Princeton University, July.
Dollar, D., Kleinberg, T. and A. Kraay, 2013, “Growth Still is Good for Poor,” Policy Research
Working Paper No. 6568, (Washington: The World Bank).
Góes, C. and D. Duque, 2016, “Como as Universidades Públicas no Brasil Perpetuam a

Desigualdade de Renda: Fatos, Dados e Soluções,” Nota de Política Pública No. 01/2016, São
Paulo: Instituto Mercado Popular.
Góes, C. and T. Matheson, 2017, “Domestic Market Integration and the Law of One Price in
Brazil,” Applied Economics Letters, Vol. 24, No. 5, pp. 84−288.
Heckman, J., 2008, “School, Skills, and Synapses,” Economic Inquiry, Vol. 46, Issue 3,
pp. 289−324, July.
International Monetary Fund, 2014, “Fiscal Policy and Income Inequality,” IMF Policy Paper
(Washington).
International Monetary Fund, 2015, Brazil: Selected Issues Paper, IMF Country Report
No. 15/122 (Washington).
Jaumotte F. and C. Osorio Buitron, 2015, IMF Staff Discussion Note, SDN/15/14 (Washington:
International Monetary Fund).

BRAZIL
Lakner, C. and B. Milanović, 2015, “Global Income Distribution: From the Fall of the Berlin Wall
to the Great Recession,” The World Bank Economic Review, Vol. 30 No. 2, pp. 203‒32.
Li, C. and J. Gibson, 2014, “Spatial Price Differences and Inequality in the People’s Republic of
China: Housing Market Evidence,” Asian Development Review, Vol. 31, No. 1, pp. 92–120.
Lustig, N., Pessino, C. and J. Scott, 2014, ”The Impact of Taxes and Social Spending on
Inequality and Poverty in Argentina, Bolivia, Brazil, Mexico, Peru, and Uruguay,” Public Finance
Review, Vol. 42, No. 3, pp. 287‒303.
Maurizio, R., 2014, “El Impacto distributive del salario minimo en la Argentina, el Brasil, Chile y
el Uruguay,” CEPAL, Serie Politicas Sociales, No. 194 (LC/L 3825).
Medeiros, M., Ferreira de Souza, P.H. and F. Avila de Castro, 2015, “A estabilidade da
Desigualdade de Renda no Brasil, 2006 a 2012: estimativa com dados do imposto de renda e
pesquisas domiciliares,” Ciência & Saúde Coletiva, Vol. 20 No.4, Rio de Janeiro, April.
Mincer, J., 1974, “Schooling, Experience, and Earnings,” NBER (Cambridge, Massachusetts:
National Bureau of Economic Research).
Neri, M., 2010, “Tackling Inequalities in Brazil, China, India and South Africa—The Role of
Labour Market and Social Policies,” OECD Publishing, Paris.
OECD, 2014, Education at a Glance 2014: OECD Indicators, OECD Publishing,

http://dx.doi.org/10.1787/eag-2014-en.
OECD, 2017, Revenue statistics in Latin America and the Caribbean, 1999‒2010.
Ostry, J.D., Berg, A. and C.G. Tsangarides, 2014, Redistribution, Inequality, and Growth, IMF
Staff Discussion Note, SDN/14/12, (Washington: International Monetary Fund).
Silveira-Neto, R. M. and C. Azzoni, 2012, “Social Policy as Regional Policy: Market and
Nonmarket Factors Determining Regional Inequality,” Journal of Regional Science, Vol. 52,
Issue 3.
Soares, F.V., Soares, S., Medeiros, M., and R. G. Osorio, 2006, “Programa de Transferência de
Renda no Brazil: Impactos Sobre a Desigualdade, Texto para Discussão,” No. 1228, Institute for
Applied Economic Research, IPEA.
Soares, S., 2012, “Bolsa Família: Its Design, Its Impacts and Possibilities for the Future,” Working
Paper No. 89, International Centre for Inclusive Growth, Institute for Applied Economic
Research (IPEA).

BRAZIL
Song, J., Price, D. J., Guvenen, F., Bloom, N. and T. von Wachter, 2015, “Firming up Inequality,”
NBER Working Paper No. 21199 (Cambridge, Massachusetts: National Bureau of Economic
Research).
Souza, P. H. G. F., 2013, “A distribuição de renda nas pesquisas domiciliares brasileiras:

harmonização e comparação entre censos, PNADS e POFS,” TD 1832, IPEA, Brasilia.
Souza, P. H. G. F. and M. Medeiros, 2013, “Diferencial salarial público-privado e desigualdade

de renda per capita no Brasil, Estudos Econômicos, São Paulo, Vol. 43, No. 1, p. 5‒28.
The World Bank, 2016, “Retaking the Path to Inclusion, Growth and Sustainability,” Report
No. 1-1431-BR, Brazil Systematic Country Diagnostic, (Washington: World Bank).
_________________, 2017, “Safeguarding Against the Reversal in Social Gains During the Economic
Crisis in Brazil, Poverty and Inequality Monitoring Note: Latin America and the Caribbean,”
(Washington: World Bank).

BRAZIL
INTEREST RATES AND INFLATION IN BRAZIL1

The conventional view among economists is that higher interest rates reduce inflation.
However, the prolonged period of low inflation and low interest rates in advanced
economies following the global financial crisis appears to be inconsistent with this view.
This sparked a debate: Do lower interest rates increase inflation (the conventional view), or
do they lead to lower inflation (the so-called Neo-Fisherian view)? This chapter finds strong
evidence in favor of the conventional view of monetary policy transmission in Brazil. While
lower inflation and lower nominal interest rates can be achieved over the long term by
targeting a lower level of inflation, this is likely to come at the cost of lower output (and
employment) in the short term—a cost that can be mitigated by enhancing monetary
policy transparency and credibility. Monetary policy transmission could be made more
efficient by reducing distortions and improving the allocation of resources in the financial
sector.
A. Introduction
1. The conventional view among economists is that higher interest rates led to lower
inflation. The rationale behind this view is that higher interest rates increase the cost of borrowing
and dampen demand across the economy, resulting in excess supply and lower inflation. In this
context, there is several channels through which higher interest rates reduce inflation, including the
exchange rate channel, the credit channel, and the bank-balance sheet channel (see Mishkin, 1996).
Here, a central bank facing the prospect of inflation being higher than its inflation target would raise
interest rates enough to increase the real (inflation-adjusted) cost of borrowing, reducing aggregate
demand and returning inflation back toward the desired level.
2. There has recently been some debate about whether lower inflation can be achieved
by setting lower policy interest rates, the so-called Neo-Fisherian effect. At the heart of the
debate, is a well-known equation in economics, the Fisher equation, which relates the nominal
interest rate to the real interest rate and expected inflation (all annualized):
Taken at face value, and assuming that the real interest rate is fixed in the long run, the equation
implies that a lower long-run inflation rate can be achieved by permanently setting the nominal
interest rate to a lower level (see Cochrane, 2016). Indeed, proponents of this view often point to the
positive relationship between nominal interest rates and inflation seen across many countries as
evidence of ‘Neo-Fisherian’ effects (see Figure 1).2
1
Prepared by Troy Matheson (WHD).
2
Neo-Fisherian effects exist in standard models used by central banks under the assumption that economic agents
have perfect foresight and do not base their decisions on the past observations. See Garcia-Schmidt and Woodford
(2015) and Garin and others (2016).

BRAZIL
Figure 1. Brazil: Headline Inflation and Policy Rate

45
40
35 Inflation (%, yoy)
30
25
20
15
10
0
Jan-98
Jan-99
Jan-00
Jan-01
Jan-02
Jan-03
Jan-04
Jan-05
Jan-06
Jan-07
Jan-08
Jan-09
Jan-10
Jan-11
Jan-12
Jan-13
Jan-14
Jan-15
Jan-16
Jan-17
Source: Haver Analytics.
3. Would a commitment to fix the policy interest rate to a lower level eventually lead to
lower inflation? To answer this question, an empirical analysis is conducted to assess the impact of
changes in Brazil’s policy rate on inflation in Section B. Section C then evaluates a simple model with
long-run Neo-Fisherian effects in the context of several countries’ historical experiences in
transitioning to lower levels of inflation. Section D concludes with a summary of key findings and
policy conclusions.
B. Empirical Analysis
4. The empirical analysis is based on Vector-Auto-Regressions (VARs). The baseline VAR is

estimated using monthly data ranging from 2003 to 2016 and contains six variables: monthly
headline inflation, the nominal interest rate (SELIC), the output gap, 12-month-ahead inflation
expectations, and monthly percent changes of commodity prices and the real-effective exchange
rate. Inflation responses to the interest rate are likely to different across different sectors of the
economy, so in addition to the baseline VAR, five additional VARs are estimated as a robustness
check, where each VAR includes inflation for a different sector of the economy. Overall, the empirical
analysis examines the impact of policy interest rate changes on headline inflation, non-regulated-
price inflation, regulated-price inflation, service-price inflation, tradable inflation, and non-tradable
inflation.3
5. Cross-correlations show that higher inflation leads to higher interest rates and higher
interest rates lead to lower inflation, consistent with the conventional view. The estimated
cross-correlation function from the baseline VAR is displayed in Figure 2; the results for all inflation
rates can be found in Appendix I, Section B. There is a statistically significant positive relationship
3
VAR lag lengths are selected using the Swartz-Bayesian Inflation Criteria. Parameter uncertainty is captured in the
analysis using bootstrapping methods, where for each VAR is resampled 1000 times.

BRAZIL
between past levels of inflation and the interest rate and a statistically significant negative
relationship between past levels of the interest rate and inflation. These results broadly reflect the
standard view of the monetary policy transmission to inflation. Since inflation tends to lead the
policy interest rate, it appears that the central bank has responded to inflation developments over
this sample, partly as the result of unanticipated demand and supply shocks (such as food and
regulated-price shocks, and exchange rate shocks). The results also suggest a peak correlation
between leads and lags of inflation and leads and lags of the interest rate of around 6 months.
Figure 2. Brazil: Correlation: Headline Inflation and

Interest Rate
(Correlation and 20th to 80th percentiles)
0.5
0.4
0.3
0.2 La g Interest Rate→Infla on
0.1
-0.1
-0.2
La g Infla on→Interest Rate
-0.3
-0.4
121110 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 101112
Month (lag) Month (l ag)
6. Structural VARs also support the conventional view that an unexpected cut in the
policy interest rate increases inflation in the short term. Responses of headline inflation to a
100 basis point cut in the policy interest are displayed in Figure 3; the results for all other inflation
rates can be found in Appendix I, Section B. Here, the uncertainty around the responses relates both
to uncertainty about the parameters of the VARs and to the recursive ordering used to identify the
monetary policy shock.4 Examining all possible recursive identification schemes allows the analysis to
be agnostic about whether an interest rate shock has a contemporaneous or a lagged impact on
inflation. The results show that an unexpected cut in the policy interest rate tends to increase
inflation over time, with the magnitude of the impact depending on the sector of the economy and
the peak impact generally occurs around 9 months after the shock. The short-term impact of a lower
interest rate on inflation is less clear cut, with identification schemes that allow for a cut in the
interest rate to immediately impact on inflation (within the same month) sometimes suggesting a
positive relationship between inflation and interest rate shocks. Overall, however, the results from
the structural VARs strongly support the standard view of monetary policy transmission.
4
Each VAR contains six variables so there is 720 different ways to order the variables to identify shocks: 69 of these
orderings lead to unique inflation responses to an interest rate shock. 1,000 parameterizations of each reduced-form
model are simulated using bootstrapping methods, leading to 69,000 different estimates of the response of inflation
to an interest rate for each VAR examined.

BRAZIL
Figure 3. Brazil: Headline Inflation after 100bps Cut in

Policy Rate (Percent, annualized, 20th and 80th percentiles)
0.80
0.60
0.40
0.20
0.00
-0.20
-0.40
-0.60
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
Months
Source: Fund staff estimates.
C. How can Lower Inflation Be Achieved over the Long Term?
7. While there is little evidence that a lower interest rate leads to lower inflation in Brazil
in the short term, the long-run Fisher equation can still help to inform policy advice. The long-
run Fisher equation is:
∗ ∗ ∗
where the steady-state nominal interest rate is equal to the steady state real interest rate plus the
inflation target. Assuming the long-run neutrality of money (i.e. nominal variables do not affect real
variables in the long run), the inflation target determines the steady state nominal interest rate. This
relationship can easily be inserted into a simple (and standard) New Keynesian model (see
Appendix I, Section C).
8. Model-based simulations show that a lower long-term inflation and nominal interest
rates can be achieved by lowering the inflation target, but this is likely to be costly in the
short run when the central bank has limited policy credibility. The results show that a reduction
in the inflation target reduces both the nominal interest rate and inflation in the long run (Figure 4).
If households and firms in the economy have expectations that are either partially forward-looking
or entirely backward looking, the transition to the new inflation target requires lower output in order
to move inflation expectations to the new target; the real interest rate must rise to reduce demand
in the short term. On the other hand, in a purely forward-looking model the central bank is fully
credible, and households and firms fully understand the future implications of monetary policy
actions and immediately embed this knowledge in their expectations. In this case, the transition of
inflation and nominal interest rates to the new target is instantaneous once the target is announced
and output is unaffected.

BRAZIL
Figure 4. Brazil: Simulated Responses to a Change in the Inflation Target

from 4.5 Percent to 2 Percent
Forward Looking Partially Forward Looking Backward Looking

(Percent) (Percent) (Percent)
20 20 20
Output Gap
15 Inflation 15 15
Policy Rate
10 10 10
5 5 5
0 0 0
-5 -5 -5
-10 -10 -10

-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112
Quarters from Inflation Targeting Announcment Quarters from Inflation Targeting Announcment Quarters from Inflation Targeting Announcment
9. Disinflation episodes across countries show that inflation was slow to adjust to lower
levels and the transition to lower inflation was costly in terms of output, reflecting
unanchored inflation expectations and limited monetary policy credibility prior to
disinflation. Figure 5 shows the behavior of inflation, interest rates, and the output gap in the two
years prior to the adoption of inflation targeting in the first five countries that formally adopted the
practice, in addition to the Volker disinflation episode in the United States, beginning in 1981.5 The
behavior of inflation, interest rates, and the output gap follow broadly similar trends across
countries. Nominal interest and inflation rates were positively correlated and tended to decline
together once the central bank formally adopted inflation targeting; output gaps generally moved
into negative territory. These results are qualitatively (and quantitatively) very similar to the
simulation results when households’ and firms’ expectations for inflation and output are not
assumed to be entirely forward looking. The large output losses during disinflation across these
countries likely reflects a high degree of inflation persistence and limited policy credibility prior to
the adoption of inflation targeting. Changing the inflation target would be likely be less costly if the
central bank had more policy credibility and more anchored inflation expectations prior to the target
change.
D. Summary and Discussion
10. There is strong evidence of the conventional view of monetary policy transmission in
Brazil, suggesting that a cut in the policy interest rate leads to higher inflation in the short
term. Cross correlations show that higher inflation leads to higher nominal interest rates and higher
interest rates result in a reduction in inflation. Since inflation tends to lead the policy interest rate
the sample examined, it appears that the central bank has responded to inflation developments,
partly as the result of unanticipated demand and supply shocks (such as food and regulated-price
shocks, and exchange rate shocks). Structural VARs also suggest that an unexpected cut in the policy
5
For each country, the output gap is defined as the percent deviation of real GDP from a linear trend.

BRAZIL
interest rate leads to a broad-based rise in inflation across the sectors examined, with the peak
impact on inflation occurring around 9 months after a monetary policy shock.
Figure 5. Brazil: Disinflation Episodes Across Countries
United States New Zealand

(Percent, annual average) (Percent, annual average)
20 20
15 15
10 10
5 5
0 0
Output Gap
-5 Interest Rate -5
Inflation
-10 -10
-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112
Quarters from Volker Disinflation (1981) Quarters from Inflation Targeting Announcment
Canada Australia
20 20
15 15
10 10
5 5
0 0
-5 -5
-10 -10
-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112
Quarters from Inflation Targeting Announcment Quarters from Inflation Targeting Announcment
United Kingdom Sweden

20 20
15 15
10 10
5 5
0 0
-5 -5
-10 -10
-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112
Quarters from Inflation Targeting Announcment Quarters from Inflation Targeting Announcment

BRAZIL
11. Model-based simulations and cross-country evidence suggest that lower inflation and
lower nominal interest rates can be achieved over the longer term if the central bank commits
to a lower inflation target. If households and firms base their output and inflation expectations on
the past (even partially), the transition to the new inflation target comes at the cost of lower output
in the short term, with larger output losses and more prolonged transition periods occurring when
expectations are more backward looking. An examination of disinflation episodes across several
countries broadly supports the key findings from model simulations that assume that expectations
were (at least) partially backward looking prior to disinflation.
12. While it appears there is no easy way to permanently lower inflation in Brazil, a lower
inflation target could be achieved at less cost with enhanced monetary policy transparency
and credibility. Enhanced credibility can better anchor inflation expectations, reduce the
persistence of inflation, improve the short run tradeoff between inflation and output, and mitigate
the cost associated should a lower inflation target be desired over the medium term. As discussed in
Domit and others (2016), there are several dimensions along which Brazil’s inflation targeting
framework can be improved to enhance transparency and credibility, including increasing the
autonomy of the central bank and changing the inflation target from a range that needs to be met
at the end of each year to a longer-term point target. The National Monetary Council made a step in
this direction in 2015 by narrowing the target range from 4.5 percent +/- 2 percent to 4.5 percent
+/- 1.5 percent from 2017.
13. Monetary policy transmission could also be made more efficient by reducing
distortions and improving resource allocation in the financial sector. There is general
agreement that the effectiveness of monetary policy in Brazil could be improved by changing
various credit policies that involve earmarking and credit subsidies. In particular, as already
proposed by the authorities, the gap between the subsidized interest rate on long-term lending (the
TJLP) and the policy interest rate (the SELIC) could be reduced over time to improve monetary policy
transmission. Linking the TJLP more tightly with the SELIC or another market-determined interest
rate (such as long-term yields on government inflation-linked debt) would enhance the transmission
of SELIC changes to longer-term interest rates. This could increase the potency of a given change in
the SELIC and contribute to lower interest rate volatility over the business cycle. Improving the
efficiency of resource allocation in the financial sector could also contribute to a lower long-term
real interest rate in Brazil, allowing for lower nominal interest rates for a given inflation target.6
6
See Minutes of the 205th Meeting of the Monetary Policy Committee (“Copom”) of the Central Bank of Brazil for a
discussion.

BRAZIL
Appendix I. Data and Robustness
A. Data
Data, Sources and Transforms
Series Source Transform*
Headline IPCA IBGE ∆log(x)*1200
Non-tradable IPCA IBGE ∆log(x)*1200
Tradable IPCA IBGE ∆log(x)*1200
Services IPCA IBGE ∆log(x)*1200
Non-regulated IPCA IBGE ∆log(x)*1200
Regulated IPCA IBGE ∆log(x)*1200
Inflation expectations (12-months ahead) BCB x
Interest rate (SELIC) BCB x-hptrend(x)
Activity Index (IBC-BR) BCB log(x)*100-hptrend(log(x))*100
Commodity Price Index (IC-BR) BCB ∆log(x)*1200
Real Effective Exchange Rate (broad) JP Morgan ∆log(x)*1200
* ∆=first difference; hptrend=Hodrick-Prescott Filter

BRAZIL
B. Empirical Results
Figure 1. Brazil: Range of Cross-Correlations Between the Interest and Inflation Rates
(Median and 20th and 80th percentiles)
Headline and Interest Rate Non-Regulated and Interest Rate

(Correlation) (Correlation)
0.5 0.4
0.4 0.3
La g Interest Rate→Infla on La g Interest Rate→Infla on
0.3
0.2
0.2
0.1
0.1
0
0
-0.1
-0.1
-0.2 La g Infla on→Interest Rate -0.2
-0.3 -0.3
-0.4 -0.4
121110 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 101112 121110 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 101112
Month (l ag) Month (l ag) Month (l ag) Month (l ag)
Regulated and Interest Rate Services and Interest Rate

0.4 0.2
La g Interest Rate→Infla on
0.15
0.3 La g Interest Rate→Infla on
0.1
0.2
0.05
0.1 0
-0.05
0
La g Infla on→Interest Rate -0.1

-0.1
-0.15
-0.2 -0.2
121110 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 101112 121110 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 101112
Tradable and Interest Rate Tradable and Interest Rate

0.4 0.3
0.3
La g Interest Rate→Infla on 0.2 La g Interest Rate→Infla on
0.2
0.1
0.1
0 0
-0.1
-0.1
-0.2
La g Infla on→Interest Rate La g Infla on→Interest Rate
-0.2
-0.3
-0.4 -0.3
121110 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 101112 121110 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 101112

BRAZIL
Figure 2. Brazil: Range of Inflation Responses to 100bps Increase in Interest Rate

(Median and 20th and 80th percentiles)
Headline Inflation Non-Regulated Inflation

(Percent, annualized) (Percent, annualized)
0.80 0.70
0.60 0.60
0.50
0.40
0.40
0.20 0.30
0.00 0.20
0.10
-0.20
0.00
-0.40
-0.10
-0.60 -0.20
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
Months Months
Regulated Inflation Services Inflation

0.80 0.30
0.60 0.20
0.40 0.10
0.20
0.00
0.00
-0.10
-0.20
-0.20
-0.40
-0.30
-0.60
-0.80 -0.40
-1.00 -0.50
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
Months Months
Tradable Inflation Non-Tradable Inflation

0.90 0.60
0.80
0.50
0.70
0.60 0.40
0.50 0.30
0.40
0.20
0.30
0.20 0.10
0.10 0.00
0.00
-0.10
-0.10
-0.20 -0.20
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
Months Months

BRAZIL
C. Simple Model
The IS curve relates the current level of the output gap, , to the lagged output gap, expectations of
the future output gap and the real interest rate (deviation from steady state):
∗
1
The Phillips curve relates the current level of inflation to inflation expectations, past inflation, and the
output gap (where 0 1 :
The monetary policy rule relates the nominal interest rate to the steady state nominal interest rate
and the expected deviation of inflation from the inflation target:
∗ ∗
where 1 to ensure a unique and stable solution, and the long-run Fisher equation is:
∗ ∗ ∗
The parameters values are 1, 0.05, 1.5, and ∗

6,The parameters in the Phillips and IS
curves related to persistence are: 1, in the forward-looking model; 0.5, in the
partially forward-looking model; and 0, in the backward-looking model.

BRAZIL
References
Cochrane, John, 2016, "Michelson-Morley, Occam and Fisher: The Radical Implications of
Stable Inflation at Near-Zero Interest Rates?" Working Paper (Hoover Institution).
International Monetary Fund, 2016, “Upgrading Brazil’s Inflation-Targeting Framework”, in

Brazil: Selected Issues, IMF Country Report No. 16/349 (Washington).
García-Schmidt, Mariana, and Michael Woodford, 2015, “Are Low Interest Rates Deflationary?
A Paradox of Perfect-Foresight Analysis”, NBER Working Paper No. 21614 (Cambridge,
Massachusetts: National Bureau of Economic Research).
Garín J, Lester and Eric Sims, 2016, “Raise Rates to Raise Inflation? Neo-Fisherianism in the New
Keynesian Model”, NBER Working Paper No. 22177 (Cambridge, Massachusetts: National
Bureau of Economic Research).
Mishkin, Frederic, 1996, “The Channels of Monetary Policy Transmission: Lessons for Monetary
Policy”, NBER Working Paper No. 5464 (Cambridge, Massachusetts: National Bureau of
Economic Research).

BRAZIL
THE EFFECT OF FISCAL CONSOLIDATION ON REAL

INTEREST RATES IN EMERGING MARKET ECONOMIES1
In 2016, the Brazilian authorities introduced reforms aimed at controlling the budget
deficit to improve fiscal discipline, boost confidence in the economy and increase growth.
In this context, the Congress passed a 20-year public spending ceiling. At the same time,
Brazil has been slowly recovering from one of the most severe recessions, during which
output fell by a cumulative 7.5 percent during 2015−16. In addition, real interest rates
have been among the highest in Emerging Market economies (EMs), further hampering
growth. Given the authorities agenda geared towards imposing fiscal discipline, we
examine the relationship between fiscal consolidation episodes and real interest rates
among EMs. The key findings suggest that there are beneficial, but relatively moderate
and short-lived effects of fiscal consolidation on real interest rates, possibly reflecting
difficulties in continuing the adjustment over a long time.
A. Fiscal Policies in Brazil
1. The policy agenda seeks to promote fiscal discipline. In December 2016, the authorities
passed a 20-year constitutional amendment limiting growth in federal noninterest spending to the
rate of consumer inflation of the previous year. To further reinforce fiscal discipline and ensure
sustainability, an ambitious social security reform bill is being discussed in Congress. Provided they
are successfully implemented, the set of policies is expected to put government accounts in order
and invigorate business confidence and growth.
2. The implementation of reforms is Figure 1. Cyclically Adjusted Primary Balance

timely for Brazil as fiscal condition have in a Set of EMs
been on a declining path. In the period (In percent of GDP)
2013−15, the debt-to-GDP ratio increased by
about 10 percentage points on average, and
the cyclically adjusted primary balance (CAPB),
i.e. the non-interest balance adjusted for
business cycle effects,2 has been consistently
worsening since 2009 (Figure 1).
3. Many observers think that fiscal

consolidation is going to reduce interest
rates. Consolidation efforts are expected by
many to have positive effects by decreasing
real short- and long-term interest rates, which Sources: IMF WEO and author’s calculations.
for Brazil have been among the highest in EMs (Figure 2). Earlier studies have shown that episodes
1
Prepared by Nina Biljanovska (WHD).
2
The methodology for the calculation of the CAPB is discussed in Section C.

BRAZIL
of fiscal tightening are likely to be accompanied by expansionary monetary policy to offset the
contractionary impact of fiscal tightening (e.g. IMF, 2010). Moreover, a successful implementation of
fiscal reforms could increase confidence, improve fundamentals and decrease the risk premium
observed in long-term interest rates. To shed light on how fiscal policies could affect interest rates in
Brazil, this study builds a panel dataset that consist of 73 countries and identifies episodes of fiscal
consolidation during 2001−15. Then, using statistical techniques, it quantifies the effect of fiscal
consolidation episodes on real short- and long-term interest rates in EMs to try to draw some
lessons that may be useful in the discussion of Brazil.3
Figure 2. Real Rates in a Set of EMs

Panel 2. Real 10-year Government Bond Yield. It is
Panel 1. Real Policy Rate. It is calculated by subtracting
calculated by subtracting the realized inflation rate from
the realized inflation rate from the nominal interest rate.
the nominal interest rate.
Sources: Haver and IMF staff calculations. Sources: Bloomberg and IMF staff calculations.
B. Fiscal Consolidation Episodes
4. The analysis employs the so-called outcome-based approach in identifying historical

cases of fiscal consolidation. Episodes are identified based on changes in the cyclically adjusted
primary balance (CAPB).4 There are different ways to calculate the CAPB.5 The approach followed
here is the one from OECD (2009):
3
Real interest rates are calculated using the Fisher equation, , where is the average annualized interest
rate in year t and is the realized inflation in the same year.
4
An alternative way to identify the fiscal consolidation episodes is the action based approach (Romer and Romer,
2010), which identifies episodes based on records of fiscal actions found in historical documents (Budget Speeches,
Budgets, central bank reports, IMF Staff Reports, IMF Recent Economic Developments reports, and OECD Economic
Surveys, see for example Guajardo, Leigh, Pescatori, 2011). Even though this method would be preferred because of
the discretionary nature of the policy actions, there is no ready-to-use dataset containing the information on such
policy changes for the large sample of countries considered in this study.
5
For example, Blanchard (1990) calculates the cyclically adjusted balance as the change between a measure of the
fiscal variable at period t, computed as a function of the unemployment rate at t-1, and the actual value of the fiscal
variable at t-1.

BRAZIL
where denotes government revenues, denotes non-interest expenditures, denotes potential

output and denotes actual output; and are the elasticities of revenues and expenditures with
respect to the output gap, which take values of 1 and 0, respectively.6 Then, an episode of fiscal
adjustment is identified using the following definition (similar to the one in Alesina and Ardagna
(2012):
A fiscal adjustment is a period of 2 years during which the cyclically adjusted primary balance/GDP
improves in each year and the cumulative improvement is at least two percentage points of the actual
GDP.
5. How many episodes of fiscal consolidation were there in our sample? Table 1 in the
Appendix summarizes the episodes of fiscal consolidation that were identified based on the
definition stated above. In total, there were 232 fiscal consolidation episodes for 73 countries in the
period 2001−15. On average, fiscal consolidations amounted to 2.2 percent of GDP per year. For the
sample of EMs, there were 35 fiscal consolidation episodes, amounting to a consolidation of
1.8 percent of GDP per year on average.7 We also try an alternative definition of an episode, defining
it as an improvement in the cyclically adjusted balance by 3 (instead of 2) percentage points of GDP.
By this criterion, there were only 18 consolidation episodes identified among EMs, not allowing for
reliable econometric analysis.
6. During the period 2001−15 no fiscal consolidation episodes were recorded for Brazil.
The public sector underwent a period of fiscal adjustment during the late 1990s, but this happened
before the start of our sample period. Throughout most of the 2000s, the cyclically adjusted primary
balance remained positive with an average of 2.8 percent of GDP for the period 2003−13. The
balance started decreasing in 2009 before turning negative in 2014.
C. Econometric Specification
7. To estimate the effects of fiscal consolidation on interest rates, the Local Projections
(LP) method proposed by Jorda (2005), was employed. In a nutshell, the advantage of this
method is that it produces Impulse Response Functions (IRFs) using ordinary least squares (OLS).
Unlike standard vector auto-regression (VAR) models, this method does not constrain the shape of
the IRF, therefore permitting a more robust lag specification. Moreover, since the IRFs are obtained
using OLS, interaction terms can easily be added to extend the analysis, allowing for cross-country
heterogeneity. The baseline specification follows Alesina and Ardagna (2012) and IMF (2010):
6
These elasticities are rough approximations in the absence of country specific information, and are likely to be more
accurate in the context of EMs.
7
The sample of EMs includes the large emerging markets based on geographical areas: Brazil, Chile, China,
Colombia, Czech Republic, Hungary, India, Indonesia, Korea, Malaysia, Mexico, Peru, Philippines, Poland, Russia,
South Africa, Thailand, Turkey.

BRAZIL
,
, , , , ∆ , ∆ , ∆ , , , ,
where denotes the real policy rate or the real 10-year government bond yield, ∆ is the
change in the cyclically adjusted primary balance, ∆ is an interaction variable between a
fiscal consolidation episode and the change in the cyclically adjusted primary balance ∗
∆ , is a vector of country fixed effects to capture country-specific characteristics, is a
vector of time fixed effects to capture global business effects, and is the error term. Finally,
∆ ,
is an interaction variable obtained by multiplying ∆ with a dummy variable that
takes the value of 1 if a country is an emerging market and zero otherwise ( ∗ ∗∆ ).
This interaction term allows for the coefficient to differ for EMs compared to all remaining
countries in the sample. Then, the marginal effect, calculated as:
,
∆ , ∗∆ ,
depicts the responses of real interest rates to fiscal consolidation episodes. Given that the marginal
effect captures how EMs respond only in comparison to other countries in the sample, we also
examine the behavior of EMs alone in Section E (not relative to other economies). In both cases the
findings are aligned.
8. Fiscal consolidation episodes are accompanied by decreases in real rates in EMs. Given
a fiscal consolidation episode, a one percentage point improvement in the change of the cyclically
adjusted primary balance is associated with approximately 100 basis points decrease in the real
policy rate (Figure 3, Panel 1). The effect is short lived, but significant for about 6 months during the
period of fiscal retrenchment. Based on the experience of EMs, this suggests that fiscal consolidation
episodes are accompanied by accommodative monetary policy to counterbalance the effects of
fiscal consolidation. The effect of fiscal tightening appears to be less pronounced on the 10-year real
government bond yield. Namely, given a fiscal consolidation episode, a one percentage point
improvement in the change of the cyclically adjusted primary balance is associated with
approximately a 70 basis points decrease in the real policy rate (Figure 3, Panel 2). The response is
only significant initially for the first couple of months during the consolidation episode. Under the
expectation hypothesis, the 10-year real rate at any point in time encompasses the average of the
future short rates over the following 10 years and a risk premium. The results suggest that fiscal
consolidation does affect both components, though it cannot be inferred whether the impact on
expectations or the risk premium dominates. Nonetheless, the effect does not seem to be long-
lasting given that the responses turn insignificantly different from zero as time progresses.

BRAZIL
Figure 3. The Response of Real Interest Rates to Fiscal Consolidation Episodes in EMs
Panel 1. Real Policy Rate Panel 2. Real 10-year Government Bond Yield
Source: IMF staff estimates. Source: IMF staff estimates.
9. One criteria to judge the success of the consolidation is by the reversals in the fiscal
adjustment, which may explain the short-lived effect of tightening on interest rates. Gupta
and others (2002) find that the probability of reversals in fiscal adjustment are as high as 70 percent
at the end of the second year (for low income countries). In line with those findings, in the sample of
countries studied here, reversals occur on average after the second year (Figure 4, Panel 1), and
almost all EMs experience reversals in the fiscal adjustment during the fourth year following the
consolidation (Figure 4, Panel 2).8 These reversals are likely to explain why fiscal tightening does not
appear to have a long-lasting effect on interest rates.
Figure 4. Fiscal Adjustment Reversals in EMs

Panel 1. CAPB reversals following a consolidation episode Panel 2. CAPB reversals four years after consolidation
1
CAPB(t+h) - CAPB|Episode
0.5
0
-0.5 1 2 3 4
-1
-1.5
-2
-2.5
All countries EM countries
Source: IMF staff calculations. Source: IMF staff calculations.
8
Fiscal adjustment reversals are calculated as the difference between the CAPB as a percentage of GDP at period t+h
and period t, where h denotes the number of periods after the fiscal consolidation episode has taken place.

BRAZIL
D. The Effect of Debt Outstanding
10. Do countries with larger outstanding debt levels experience larger benefits from fiscal
consolidation? There is evidence that the response of interest rates to movements in fiscal policy
tends to be of larger magnitude in countries that exhibit higher outstanding debt levels (e.g.
Nakamura and Yagi, 2015 among others). To examine this question, we augment the above
equation by an additional term to capture the difference in response of interest rates in emerging
markets with different outstanding debt levels. In this case, the above equation becomes
, , ,
, , , , ∆ , ∆ , ∆ , ∆ ,
, , ,
where ∆ , ,
is an interaction variable between ∆ ,
and a dummy variable, which
takes a value of one if a country has a gross outstanding debt higher than 60 percent of GDP and
zero otherwise.9
11. The negative relationship between interest rates and episodes of fiscal consolidation is
observed also when controlling for outstanding debt levels. Given a fiscal consolidation episode,
a one percentage point improvement in the change in the cyclically adjusted primary balance
reduces the interest rate by almost 200 basis points (Figure 5, Panel 1). The finding is qualitatively
the same as before, but quantitatively stronger. On the other hand, the response of the real 10-year
government bond yield is of the same magnitude as before (around 70 basis points). However, the
effect appears to be significant only during the first year while the fiscal consolidation is still
ongoing (Figure 5, Panel 2).
(D>60 percent of GDP)
9
This threshold level, although arbitrary, is familiar owning to the Maastricht Treaty.

BRAZIL
E. Robustness
12. Control for real GDP-per-capita. The country fixed effects do not control for time-varying
country specific characteristics. Therefore, we augment the equations above by the GDP-per-capita
levels. The main results are not affected by the addition of real GDP since the coefficients on our key
variables of interest, depicting the marginal effect of fiscal consolidation episodes on interest rates,
remain intact. This outcome suggests that the country and time fixed effects capture the cross-
country and time variation and the GDP variable is redundant in the specification.
13. Control for terms of trade. Since the cyclically adjusted balance may be influenced by
other factors such as the terms of trade, especially for countries that are commodity exporters, we
test additional specifications by including countries’ terms of trade (defined as the change in the
index) as a control. The coefficient on the terms of trade variable is not significant, suggesting as
before that the country and time fixed effects control for the terms of trade variation.
14. Change the definition of fiscal consolidation episodes. To examine the extent to which
the results are sensitive to the rate at which a fiscal consolidation episode is defined, we modify the
definition of an episode to be an improvement in the cyclically adjusted balance by 1.5 (instead of 2)
percentage points of GDP. In this case, the effect of fiscal consolidation on interest rates is
qualitatively the same, but the responses are of somewhat lower magnitude. This result is not
surprising given the more moderate degree of fiscal tightening.
15. Restricting the sample to EMs countries. The earlier analyses reflect how the responses of
EMs differ from the rest of the countries considered in the sample. Here, we inspect whether the
same specification delivers similar conclusions on the response of interest rates to fiscal
consolidation episodes when the sample is restricted only to EMs. The regression equation now
excludes the interaction term ∆ ,
and the coefficient of interest is . Even though the
sample now is smaller, the key results continue to hold. The real policy rate responds negatively to a
fiscal consolidation episode, but the magnitude is somewhat lower (Figure 6, Panel 1). Similarly, the
real government bond yield responds negatively to a fiscal consolidation episode with a somewhat
lower magnitude (Figure 6, Panel 2). This difference in responses between the full sample and the
restricted sample suggests that not only EMs respond differently to all other countries, but the
negative relationship between real interest rates and fiscal consolidation episodes holds within the
EM sample alone.

BRAZIL
(Restricted sample to EMs only)
F. Conclusion and Policy Recommendations
16. Fiscal consolidations tend to have beneficial, but relatively moderate and short-lived
effects on real interest rates. The study shows that in response to a fiscal consolidation episode,
the real short-term rate is likely to decrease by approximately 100 basis points, whereas the real
bond yield is expected to decrease by about 70 basis points. The effects, however, appear to be
significant only for a short period, likely reflecting the proclivity of fiscal adjustments to falter and
suffer reversions after a period of time.
17. Efforts to impose fiscal discipline should be accompanied by policies aimed at

alleviating the inefficiencies present in the financial sector. Based on this analysis, the fiscal
policies implemented and those on the agenda are certainly in the right direction to decrease rates
and allow for more stable growth in the long-run. However, fiscal problems are not the sole culprit
of high rates, implying that other factors such as inefficiencies present in the financial sector could
provide further explanation to the behavior of interest rates observed in Brazil.
18. What does this mean for Brazil? In principle, we could expect some positive effects from
fiscal consolidation, but it is difficult to draw definitive conclusions because of the peculiarities.
Brazil’s episode of fiscal consolidation will be long, but it has been preannounced, and is also
accompanied by monetary easing. The main lesson to draw is that gains in interest rates can be
expected if fiscal consolidation can be maintained.

BRAZIL
Appendix I. Data
Table 1 contains the list of all the countries used in the analysis and the years for which fiscal
consolidation episodes have been identified using the methodology discussed in the main text.
Table 1. Outcome-Based Approach: Episodes of Fiscal Consolidation

Economy Fiscal Consolidation
Albania 2002 2003
Algeria 2005 2006
Argentina 2003 2004 2012 2013
Australia 2012 2013
Austria
Belgium
Bosnia and Herzegovina 2002 2003 2004 2005 2006 2009 2010 2011
Brazil
Bulgaria 2003 2004 2011 2012
Canada
Chile 2003 2004 2010 2011
China
Colombia 2001 2011 2012
Costa Rica
Croatia 2012 2013 2014 2015
Cyprus 2001 2004 2005 2006 2007 2012 2013 2014
Czech Republic 2003 2004 2010 2011 2013
Denmark 2003 2004 2013 2014
Dominican Republic 2001 2002 2005 2013 2014 2015
Ecuador 2006 2007 2010 2011
Estonia 2001 2002 2003 2004 2009 2010
Finland
France
Georgia 2001 2004 2010 2011
Germany 2011 2012
Greece 2005 2010 2011 2012 2013
Guatemala
Hungary 2003 2004 2007 2008 2009 2012
Iceland 2003 2004 2005 2009 2010 2011 2012 2013 2014
India 2004 2005 2006 2007
Indonesia 2001 2002
Ireland 2003 2004 2011 2012 2013 2014
Israel 2003 2004 2005
Italy 2011 2012
Jamaica 2003 2012 2013

BRAZIL
Table 1. Outcome-Based Approach: Episodes of Fiscal Consolidation (Concluded)

Economy Fiscal Consolidation
Japan 2004 2005 2013 2014 2015
Jordan 2003 2004 2014 2015
Kazakhstan 2005 2006 2009 2010 2011
Korea
Latvia 2001 2003 2004 2009 2010
Lebanon 2001 2002 2003 2005 2006 2010 2011 2014
Lithuania 2001 2002 2003 2009 2010 2012 2013 2014 2015
Malaysia 2001 2002 2004 2005
Mexico
Morocco 2006 2007 2013 2014
Netherlands 2004 2005 2012 2013
New Zealand 2002 2003 2011 2012 2013
Norway 2004 2005 2006
Pakistan 2002 2003 2013 2014
Panama 2005 2006 2007
Peru 2005 2006 2007 2010 2011
Philippines 2005 2006
Poland 2011 2012
Portugal 2011 2012 2013
Romania 2009 2010 2011 2012
Russia 2001 2003 2004 2005 2010 2011
Serbia 2003 2004 2005 2014 2015
Slovak Republic 2003 2011 2013 2014
Slovenia 2012 2014 2015
South Africa
Spain 2012 2013 2014
Sri Lanka 2001 2002 2003
Sweden 2004 2005
Switzerland
Taiwan Province of China
Thailand 2003
Turkey
Ukraine 2002 2003 2005 2006 2009 2010 2013 2014 2015
United Kingdom 2010 2011
United States 2012 2013 2014
Uruguay 2000 2001 2002 2003
Venezuela 2002 2003 2004 2005 2006 2011 2013 2014
Vietnam 2014 2015

BRAZIL
The data span over the period 2001−15, however the panel is unbalanced due to availability prior to
2002 with respect to the fiscal variables for some of the countries. All the variables along with the
data sources are listed in Table 2. The series for GDP per capita were converted to real by using the
countries’ respective GDP deflators from the WEO database. Potential output was estimated using
the HP-filter with a penalizing parameter set to 6.5, which corresponds to yearly data.
The sample of Emerging Market economies (EMs) includes the largest EMs based on geographical
area: Brazil, Chile, China, Colombia, Czech Republic, Hungary, India, Indonesia, Malaysia, Mexico,
Peru, Philippines, Poland, Russia, South Africa, Korea, Thailand, Turkey.
Table 2. Data Sources

Variable Source
Real GDP (total) WEO Database
Real GDP (per capita) WEO Database
Government revenues (general) WEO Database
Government expenditures (general) WEO Database
Government interest payments, staff calculations WEO Database
Government debt as a percentage of GDP WEO Database
Policy interest rate National authorities (Haver), Bloomberg Financial
Markets
10-year government bond yield Bloomberg Financial Markets

BRAZIL
Appendix II. Methodology

The Local Projection (LP) method, suggested by Jorda (2005), comprises of projecting the future
behavior of the dependent variable (real interest rates in this study) on an unchanged set of
explanatory variables. Then, a series of regressions are estimated in which the set of explanatory
variables remains unchanged and only the dependent variable is being forwarded one period ahead
for every regression. The coefficients of interest, depicting marginal effects, are stored and plot as
impulse responses. More specifically, for the regression equation in Section C, the marginal effects
of a fiscal consolidation episode on interest rates are given by
,
∆ , ∗∆ , .
In this way, the coefficients above trace the response of interest rates at t+h to a fiscal consolidation
episode that has occurred at time t.
One of the challenges that this estimation poses is the serial correlation due to lagging and
successive leading of the dependent variable. Furthermore, it assuming cross-sectional
independence of the error terms might be problematic. To control for both serial correlation and
cross-sectional dependence we use Driscoll-Kraay standard errors.1
1
Driscoll and Kraay (1998) suggest an estimator that generates heteroskedasticity and autocorrelation-consistent
standard errors that are also robust to cross-sectional dependence (see for e.g. Caselli and Roitman, 2015).

BRAZIL
References
Alesina, Alberto, and Silvia Ardagna, 2012, “The Design of Fiscal Adjustments,” Tax Policy and
the Economy, Vol. 27.
Blanchard, Olivier J., 1990, "Suggestions for a New Set of Fiscal Indicators," OECD Economics
Department Working Papers No. 79.
Caselli, Francesca G. and Agustin Roitman, 2015, “Non-Linear Exchange Rate Pass-Through in
Emerging Markets,” IMF Working Paper No. 16/1 (Washington: International Monetary Fund).
Driscoll, John and Aart Kraay, 1998, “Consistent Covariance Matrix Estimation with Spatially
Dependent Panel Data,” The Review of Economics and Statistics, Vol. 80(4), pp. 549-560.
Guajardo, Jaime, Daniel Leigh, and Andrea Pescatori, 2011, “Expansionary Austerity: New
International Evidence,” IMF Working Paper No. 11/158 (Washington: International Monetary
Fund).
Gupta, Sanjeev, Benedict Clements, Emanuele Baldacci, and Carlos Mulas-Granados, 2002,
"Expenditure Composition, Fiscal Adjustment, and Growth in Low-Income Countries," IMF
International Monetary Fund, 2010, “Will it Hurt? Macroeconomic Effects of Fiscal

Consolidation,” World Economic Outlook, Chapter 3.
Jorda, Oscar, 2005, “Estimation and Inference of Impulse Responses by Local Projections,”
American Economic Review, Vol. 95(1), pp. 161−82.
Nakamura, Koji and Tomoyuki Yagi, 2015. “Fiscal Consolidation and Long-term Interest Rates,”
Bank of Japan Working Paper, No. 15-E-10.
OECD, 2009, “Computing Cyclically Adjusted Balances and Automatic Stabilizers,” Fiscal Affairs
Department.
Romer, Christina and David Romer, 2010, “The Macroeconomic Effects of Tax Changes:
Estimates Based on a New Measure of Fiscal Shocks,” American Economic Review, Vol. 100,
pp. 763−801.

BRAZIL
BOOSTING PRODUCTIVITY: TAXES AND RESOURCE

MISALLOCATION IN BRAZIL1
A. The Productivity Challenge in Brazil
1. Once Brazil recovers from its recent economic crisis, raising productivity will remain a
top priority challenge. Brazil has struggled for several decades to generate strong and sustained
productivity growth, the key driver of living standards over the long term. Brazil's productivity grew
by only 1.6 percent a year between 2001 and 2013 (Figure 1). By comparison, China and Russia, also
considered emerging economies, saw their productivity rates rise by 9.6 and 3.5 percent,
respectively (Ceratti, 2016).
Figure 1. Labor Productivity Growth in Selected Countries
Source: World Bank
2. Brazil’s productivity malaise can be explained by multiple factors, including significant

resource misallocation. At the macroeconomic level, weak productivity growth can be explained by
multiple factors, such as a high cost of finance and doing business, a poor state of the country’s
physical infrastructure, limits to competition resulting from domestic regulation, trade barriers, and
inadequate skills of the labor force (World Bank, 2016). At the microeconomic level, lack of total
1
Prepared by Henrique Fialho Barbosa (WHD) and Carlos Mulas-Granados (FAD), based on the IMF’s Fiscal Monitor
chapter on the same topic (IMF, 2017).

BRAZIL
factor productivity growth reflects weak productivity performance by individual firms, which is
typically related to the poor use of existing resources (labor and capital) within the country.
3. Reducing resource misallocation is a multidimensional task which requires upgrading

the tax system. Resource misallocation is often the result of many poorly designed economic
policies and market failures that prevent the expansion of efficient firms and promote the survival of
inefficient ones. Reducing misallocation therefore requires the use of multiple policy levers,
including structural (financial, labor, and product market) reforms which have proved successful in
improving resource allocation efficiency.2 This appendix makes the case that upgrading the tax
system in Brazil is also key to boosting productivity by reducing distortions that prevent resources
from going to where they are most productive.
B. Resource Misallocation in Brazil
4. Resource misallocation reflects a poor distribution of resources across firms, reducing

the total output that can be obtained from existing capital and labor. Resource misallocation is
apparent in the dispersion in revenue productivity levels across firms, even within narrowly defined
industries that produce similar goods. When dispersion is wide, reallocating resources from firms
with low revenue productivity to firms with high revenue productivity increases output, simply by
using the same resources differently. For example, imagine a situation in Brazil with two firms within
the same industry that have identical technologies but face different tax treatments. This could
happen if an inefficient small firm benefits from the tax exemptions under the SIMPLES tax regime,
which is available to companies below a certain revenue threshold. In that case, the larger efficient
company, using the economies of scale generated by its more efficient use of resources, would be
fully taxed. In this scenario, aggregate output would be higher if capital were reallocated toward the
more efficient firm, resulting in more investment in higher return projects. However, some resources
would remain employed in inefficient firms that could afford to undertake investments in projects
with lower pre-tax returns and survive thanks to lower taxation.
5. Brazil suffers from significant resource misallocation, as captured by high dispersion

of revenue productivities across firms. Borrowing the data and the methodology from the IMF’s
Fiscal Monitor (IMF, 2017), Figure 2 illustrates the dispersion in firm-level revenue productivities in
Brazil and the top performing country in the sample.3 The figure shows that dispersion in the more
efficient country is much narrower than that observed in Brazil. Consequently, Brazil would be able
to reap substantial gains if it were able to move resources from those firms with lower productivity
(on the left tail of the distribution) to those with higher revenue productivity (those in the right tail).
2
For selected advanced economies, Dabla-Norris and others (2015) show that TFP gains from improving factor
allocation across sectors average about 9 percent.
3
According to Hsieh and Klenow (2009), "revenue productivity (the product of physical productivity and a firm's
output price) should be equated across firms [in the same industry] in the absence of distortions. To the extent
revenue productivity differs across firms, we can use it to recover a measure of firm-level distortions."

BRAZIL
Figure 2. Brazil: Distribution of Firm-Level Revenue Productivities 1/
35
30
25
20
Percent of firms
15
10
0
-3.4-3.0-2.6-2.2-1.8-1.4-1.0-0.6-0.2 0.2 0.6 1.0 1.4 1.8 2.2 2.6 3.0 3.4
Log of firm revenue productivity scaled by corresponding
country-industry average
Top performer BRA

1/ Data from the Fiscal Monitor (April 2017). The figure shows the distribution
for firms in the manufacturing sector for each country. Top performer is
defined as a country at the 90th percentile of the cross-country distribution
of the aggregate resource allocation efficiency index for manufacturing.
6. In a fully efficient and frictionless equilibrium, Brazil could potentially increase its total
factor productivity by nearly 60 percent. If marginal products were equalized within each
industry, industry-level TFP would be at an efficient level. The Resource Allocation Efficiency ratio
(RAE) is the ratio of actual TFP to this efficient TFP, aggregated across sectors.4 Figure 3 plots the
Resource Allocation Efficiency (RAE) for all 67 countries in the sample.5 It shows that Brazil is in the
bottom third of the distribution. Implementing reforms to help Brazil to transition from its current
level of efficiency to the level of the country in the 90th percentile of the distribution (in red) could
increase Brazil’s total factor productivity by 57.3 percent. These estimates of TFP gains present a
4
Using a Cobb-Douglas aggregator.
5
Using firm-level data for 10 advanced economies (from ORBIS database) and 57 developing countries (from World
Bank Enterprise Surveys), the Fiscal Monitor (IMF, 2017) calculates resource allocation efficiency for each industry
within each country.

BRAZIL
lower bound. While within-industry misallocation accounts for the bulk of cross-country productivity
differences, reducing misallocation across broad economic sectors can also spur aggregate
productivity and output. Although these numerical estimates should be treated with caution, as they
are subject to statistical error like all estimates, they strongly suggest that Brazil could stand to gain
much if it could take steps to improve resource allocation and productivity.
Figure 3. Brazil: Cross-Country Resource Allocation Efficiency 1/

90
Resource allocation efficiency (actual sector TFP as a percent of efficient
80
70
60
Brazil, 2009
level of sector TFP)
50
40
30
20
10
Source: IMF Fiscal Monitor, April 2017

1/ Estimates of resource allocation efficiency follow Hsieh and Klenow (2009). Estimates
correspond to the latest available data for manufacturing sector.
7. Potential TFP gains from reducing resource misallocation could lift annual real GDP
growth rate by roughly 2 percent of GDP. By comparing the gains in TFP and, consequently, the
gains in output that could result from some interventions to reduce specific distortions, it is possible
to estimate the GDP growth effects from such interventions. Assuming a transition path of 20 years,
reducing resource misallocation (by moving Brazil to the 90th percentile of the distribution of
resource allocation efficiency) translates into a higher annual real GDP growth rate of 2.29 percent
for Brazil.6 One way to read this estimate is that when one considers the distance in output per
capita between Brazil and the group of more developed/productive countries today, a substantial
6
We assume a transition period of 20 years to make our results comparable to those calculated for other countries in
the world (see IMF, 2017).

BRAZIL
fraction could be made up just by improving the way Brazil uses the resources it already has. Of
course, moving resources from one part of the economy to another is a slow process, hence the
assumption of a 20-year period for catching up to the higher standard of resource allocation
efficiency. In any case, as shown in Figure 4, these gains would be much higher than average gains
for emerging market economies or the five largest economies in Latin America (LA5), partially
because Brazil starts further behind in terms of allocation efficiency.
Figure 4. Brazil: Growth Impact from Improving Resource Allocation Efficiency 1/

2.5
2.0
1.5
Real GDP growth
1.0
0.5
0.0
Brazil EMEs LIDCs LA5
Sources: April 2017 Fiscal Monitor.
1/ The figure shows medians across country groups. Estimates are computed
based on the assumption that the other sectors could achieve TFP gains similar to
those estimated for the manufacturing sector and that there are no adjustment
costs. EMEs = emerging market economies; LIDCs = low-income developing
countries; LA5 = Median for Brazil, Chile, Colombia, Mexico, Peru.
C. Upgrading the Tax System to Reduce Resource Misallocation
8. Among other factors, upgrading the tax system is key for reducing resource
misallocation and increasing productivity and growth. Several policies may be behind high levels

BRAZIL
of resource misallocation. For example, Restuccia and Rogerson (2013, 2016) survey the literature
and point to legislated provisions that vary by firm characteristics (for example provisions of the tax
code that vary with size, tariffs applied to particular goods, employment protection measures, or
product market regulations that limit market access); discretionary provisions made by the
government that favor specific firms (for example subsidies, selective tax enforcement, or
preferential loans); and market imperfections (for example monopoly power or incomplete financial
markets). For Brazil, De Vries (2014) shows that taxes, regulation and access to credit create
distortions to output and capital that vary by firm size, thus generating resource misallocation. Appy
(2017) explains that the tax system in Brazil generates allocative distortions by altering the relative
prices of labor and capital, and by inducing a suboptimal allocation of production by sectors and
regions.
9. Distortive tax policy and tax administration are among the important factors that
policymakers need to bear mind when tackling the productivity challenge in Brazil. 11 Below,
we examine a selection of tax policies to explore the channels through which they generate resource
misallocation in Brazil, and provide some estimates for potential gains, based again on the
methodologies outlined in the Fiscal Monitor (IMF, 2017). The analysis includes taxes that
discriminate across capital asset types (leading to differentiated treatment of firms as the propensity
to use the various asset types as input varies across firms), across firm characteristics such as their
sources of financing (debt or equity), and their degree of informality.7
Tax Disparity
10. When disparities in effective tax rates across asset types push investors toward lower-return,
tax-favored projects, resource misallocation increases. One way in which the tax code introduces
distortions in the economy is by differential taxation across factors of production. In this situation,
firms will choose their factor allocations partly based on tax concerns, displacing productive
concerns. Also, firms that are factor-intensive on the privileged factor can afford being less
productive and still be competitive due to their tax advantage. A standard method to estimate
resource misallocation driven by tax disparity is to compare effective marginal tax rates (EMTRs)
across various types of assets.8 The significant variation in EMTRs for various asset types in Brazil and
half of the countries shown in the Fiscal Monitor (IMF, 2017) arises from differences between the
rates at which the tax code allows businesses to deduct the cost of assets (known as tax
depreciation) and the rates at which those assets actually wear out or become obsolete (economic
depreciation).
7
Because industry or firm specific tax policy indicators are not readily available for a wide set of countries, like in the
Fiscal Monitor (IMF, 2017) most of the analyses in this section rely on a difference-in-differences approach to assess
the effect of tax distortions on resource misallocation.
8
The EMTR on capital income measures the tax burden applied to before-tax capital income realized over an
investment’s lifetime, implied by major provisions of the corporate tax code. These major provisions include the
statutory federal tax rates, surcharges, local tax rates, depreciation rates and accelerated depreciation, treatment of
inventories, and interest deductibility.

BRAZIL
11. Reducing resource misallocation associated to tax disparity in Brazil can increase
productivity and growth significantly. The differential taxation between machinery and buildings
could be used an indirect measure of the differential taxation between capital-intensive and labor-
intensive industries leading to resource misallocation. Brazil suffers from the effects of high tax
disparity, meaning that there are higher effective taxes for machines than for buildings. This explains
why the country tends to have a lower share of machinery, compared to those countries with lower
tax disparity. Eliminating most of the existing tax disparity and bringing Brazil to the level of tax
disparity existing at the 90th percentile of the distribution, would raise resource allocation efficiency
in Brazil by 12.3 percentage points and increase annual GDP by 0.58 percentage points (Figure 5).
This potential improvement would be four times larger than the potential gain for emerging
economies and three times the gains for the group of Latin American economies.9
Figure 5. Brazil: GDP Growth Effects from Reducing Tax Disparity 1/

0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
Brazil Chile Colombia Mexico Peru Median LatAm Emerging
1/ The figure includes medians across groups of Latin American and Emerging
countries.
9
Note that underlying data for Brazil used in the Fiscal Monitor (IMF, 2017) is from 2009 as reported by the World
Bank Enterprise Survey. This is when still Brazil was still among the best performing emerging markets, a situation
which has changed since 2014 due to one of the deepest recessions in its history. Therefore, inefficiencies may be
larger now and gains from reducing resource misallocation could be even higher today.

BRAZIL
The group of Latin American countries includes Argentina, Belize, Bolivia, Chile, Colombia, Costa
Rica, Ecuador, El Salvador, Falkland Islands, French Guiana, Guatemala, Guyana, Honduras, Mexico,
Nicaragua, Panama, Paraguay, Peru, Suriname, Uruguay and Venezuela. The group of Emerging
countries includes Chile, China, Colombia, Czech Republic, Egypt, Greece, Hungary, India, Indonesia,
South Korea, Malaysia, Mexico, Peru, Philippines, Poland, Qatar, Russia, South Africa, Taiwan,
Thailand, Turkey, and United Arab Emirates.
Corporate Debt Bias
12. Giving tax breaks to debt but not to equity financing of risky projects increases
resource misallocation. Governments can introduce distortions by privileging some financing
methods over others. Corporate debt bias occurs when firms are allowed to deduct interest
expenses, but not returns to equity, in calculating corporate tax liability, raising the cost of equity
financing compared to debt financing. Many young companies, especially innovative ones and
R&D-related startups finance their operations with equity due to the risky nature of their businesses
because there are no collateral requirements and shareholders share in upside returns (Stiglitz, 1985;
Hall, 2002; Brown and others, 2009). Therefore, corporate debt bias induced by the tax system not
only distorts the financing choice, but it can also create resource misallocation by imposing a higher
marginal tax on R&D investment compared to other capital spending.
13. Reducing the corporate debt bias can help increase productivity and growth
associated to innovative companies. The debt bias can be measured as the EMTR on equity minus
the EMTR on debt.10 The results from the Fiscal Monitor (IMF, 2017) show that R&D intensive
industries—which are more exposed to debt bias—have lower resource allocation efficiency in
countries where debt bias is higher. Using these estimates, we calculate that if Brazil were to reduce
its corporate debt bias to the level observed in the 10th percentile of the sample (thus moving the
country to the 90th percentile of the distribution of resource allocation efficiency), it could raise its
resource allocation by 5.56 percentage points and annual GDP growth in Brazil by 0.27 percentage
points (Figure 6). These gains would be 40 percent higher than for the average of Latin American
countries and almost three times larger than for the average of Emerging countries.
10
Note that in the case of debt-financed investment, the combination of interest deductions and accelerated
depreciation can exceed taxes paid on the associated income, resulting in negative EMTRs.

BRAZIL
Figure 6. Brazil: GDP Growth Effects from Reducing Corporate Debt Bias 1/
0.30
0.25
0.20
0.15
0.10
0.05
0.00
-0.05
countries.
Tax Informality
14. Tax informality is not only a problem for revenue collection but also for productivity.
The literature defines informal firms are those that fail to pay the full amount of tax due, recognizing
that there are many reasons why a firm or individual might not pay taxes (Kanbur and Keen, 2014).
Noncompliance with taxes reduces productivity by interfering with firm entry and exit. Informal firms
enjoy a relative cost advantage over their tax compliant competitors through tax evasion and
circumvention of regulations. This allows informal firms to stay in business despite their low
productivity, increasing their weight in the economy at the expense of more productive firms.
15. Reducing tax informality in Brazil can help increase productivity and growth further.
Results from the Fiscal Monitor (IMF, 2017) suggest that, as tax enforcement improves and the
prevalence of informality falls among registered firms, less productive firms will exit the market,
releasing resources that can then migrate to more productive firms. Reducing tax informality to the
level of the best performers can increase productivity in Brazil by 2.58 percentage points and annual
GDP by 0.12 percentage points (Figure 7). While significant, these gains would be smaller in Brazil

BRAZIL
than in the groups of Latin American and Emerging countries, reflecting the improvements that
Brazilian tax administration has experienced in the last decades.
Figure 7. Brazil: GDP Growth Effects from Reducing Tax Informality 1/

0.20
0.18
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
countries.
D. Conclusions and Policy Recommendations
16. Resource misallocation in Brazil is significant. Resource misallocation arises from

distortions that prevent the expansion of efficient firms and promote the survival of inefficient ones.
It is apparent in a wide dispersion in productivity levels across firms, even within narrowly defined
industries.
17. Brazil can reap substantial TFP gains from reducing resource misallocation. By enabling
firms to catch up with the productivity frontier, Brazil can reduce resource misallocation. Using
methodologies developed in the Fiscal Monitor, it is possible to estimate that reducing resource

BRAZIL
misallocation could increase productivity by more than 50 percentage points in Brazil. In other
words, the evidence suggests that with existing resources, and simply by improving the way these
are used, output could be that much higher. If one could imagine that a reallocation of resources
could be implemented gradually over 20 years in such a way that Brazil’s levels of efficiency could
catch up with those of leading countries, that would boost annual GDP growth by more than
2 percentage points during this catch-up period. While these estimates are subject to statistical
uncertainty, like any estimates of this type, they are suggestive of the scope for improvement that is
available to Brazil if it can focus on a productivity-increasing agenda.
18. Resource misallocation can be reduced by upgrading the tax system. Distortions are
created by several factors, like poorly designed economic policies and market failures. The tax
system is also an important factor. Countries, such as Brazil, can reap TFP gains by upgrading the
design of their tax system, to ensure that firms’ decisions are made for business reasons and not tax
reasons. Significant gains can also be achieved if Brazil addresses tax treatments that discriminate by
asset type, sources of financing, or by firm characteristics such as tax informality. The estimates
shown in the preceding sections suggest that by addressing tax-related distortions, Brazil could
bridge about half of the gap in productivity relative to leading countries. Again, exact estimates of
potential gains need to be taken with caution; but they confirm that there are important benefits for
the country that can be reached by pursuing a well-designed, productivity-oriented tax reform.

BRAZIL
References
Appy, B., 2017, “Tributação e produtividade no Brasil,” In Bonelli and others Anatomia da
Produtividade no Brasil, Rio de Janeiro: FGV-IBRE.
Ceratti, M., 2016, “Productivity: A Key Issue for Fighting Poverty in Brazil,” World Bank News:
http://www.worldbank.org/en/news/feature/2016/08/24/brazil-increase-productivity-share-
prosperityCeratti (Washington).
Brown, J. R., S. M. Fazzari, and B.C. Petersen, 2009, “Financing Innovation and Growth: Cash Flow,
External Equity, and the 1990s R&D Boom,” The Journal of Finance, Vol. LXIV, No. 1, February.
Dabla-Norris, E., S. Guo, V. Haksar, M. Kim, K. Kochhar, K. Wiseman, and A. Zdzienicka, 2015, “The
New Normal: A Sector-Level Perspective on Productivity Trends in Advanced Economies,” IMF Staff
Discussion Note (March) (Washington: International Monetary Fund).
De Vries, G., 2014, “Productivity in a Distorted Market: The Case of Brazil’s Retail Sector,” The Review
of Income and Wealth, Vol. 60(3).
Hall, B. H., 2002, “The Financing of Research and Development,” Oxford Review of Economic Policy,
Vol. 18, No. 1.
Hsieh, C.-T., and P. Klenow, 2009, “Misallocation and Manufacturing TFP in India and China,”
Quarterly Journal of Economics, Vol. 124(4), pp. 1403–448.
Kanbur, R. and M. Keen, 2014, “Thresholds, Informality and Partitions of Compliance,” International
Tax and Public Finance, Vol. 23, pp. 536–59.
International Monetary Fund, 2017, “Fiscal Monitor: Achieving More with Less Together,” (Chapter 2:
Boosting Productivity: How Can Tax Policy Reduce Resource Misallocation?).
http://www.imf.org/en/Publications/FM/Issues/2017/04/06/fiscal-monitor-april-2017, April.
Restuccia, D. and R. Rogerson, 2013, “Misallocation and Productivity,” Review of Economic Dynamics
Vo. 16, pp. 1–10.
Restuccia, D. and R. Rogerson, 2016, “The Causes and Costs of Misallocation,” NBER Working Paper
No. 23422 (Cambridge, Massachusetts: National Bureau of Economic Research).
Stiglitz, J.E., 1985, “Credit Markets and the Control of Capital,” Journal of Money, Credit and Banking,
Vol. 17, No. 2 (5), pp. 133‒52.
World Bank, 2016, Brazil—Systematic Country Diagnostic,

http://documents.worldbank.org/curated/en/180351467995438283/Brazil-Systematic-country-
diagnostic (Washington: World Bank Group).

Articulo 5

Uploaded by

Copyright:

Available Formats

Articulo 5

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Articulo 5

Uploaded by

Copyright:

Available Formats

IMF Country Report No.

Copies of this report are available to the public from

International Monetary Fund • Publication Services

International Monetary Fund

© 2017 International Monetary Fund

Approved By Prepared by H. Barbosa, C. Góes, N. Biljanovska, I. Karpowicz,

INVESTMENT IN BRAZIL: FROM CRISIS TO RECOVERY______________________________ 5

BRAZIL’S BUSINESS ENVIRONMENT AND EXTERNAL COMPETITIVENESS _______ 19

D. External Competitiveness ___________________________________________________________ 26

DISTRIBUTIONAL EFFECTS OF BRAZIL’S PENSION REFORM _______________________ 37

INEQUALITY IN BRAZIL: A MICRO-DATA ANALYSIS _______________________________ 49

2 INTERNATIONAL MONETARY FUND

3. Income Inequality in Brazilian States: Some Evidence of Convergence ______________ 54

INTEREST RATES AND INFLATION IN BRAZIL ______________________________________ 74

THE EFFECT OF FISCAL CONSOLIDATION ON REAL INTEREST RATES IN

INTERNATIONAL MONETARY FUND 3

BOOSTING PRODUCTIVITY: TAXES AND RESOURCE MISALLOCATION IN

References ____________________________________________________________________________ 110

4 INTERNATIONAL MONETARY FUND

INVESTMENT IN BRAZIL: FROM CRISIS TO

consumption, and widespread optimism about Brazilian 105

2. Some of the factors that are hampering

years, Brazil’s terms of trade stabilized and improved in early 120

2016. By reducing funding costs and increasing profitability, 70

a prolonged recession that has damaged Brazil’s long-term

INTERNATIONAL MONETARY FUND 5

Gross Fixed Capital Formation

Source: Staff estimates and Haver Analytics.

6 INTERNATIONAL MONETARY FUND

B. Aggregate Empirical Analysis

INTERNATIONAL MONETARY FUND 7

where ∆ represents the quarterly change in each variable.

6. The estimated model provides a good description of behavior of investment over

8 INTERNATIONAL MONETARY FUND

Figure 2. Brazil: Decomposition of Gross Fixed Capital Formation Growth

C. A Closer Look at Investment at the Firm Level

 Tobin’s Q to capture the expected marginal return on investment.

INTERNATIONAL MONETARY FUND 9

10. The baseline model specification is:

10 INTERNATIONAL MONETARY FUND

Table 1. Brazil: Variable Definitions

Change in Economic Policy Uncertainty by Scott Baker, Nicholas Bloom

INTERNATIONAL MONETARY FUND 11

Table 2. Brazil: Regression Results, All Sectors

Number of observations 10,763 10,378 10,144 9,805 2,392 2,311 2303

D. Summary and Conclusion

12 INTERNATIONAL MONETARY FUND

Table 3. Brazil: Regression Results, by Sector

Consumer Consumer Real Telecomm-

INTERNATIONAL MONETARY FUND 13

Appendix I. Data, Results, and Robustness

A. Economic Policy Uncertainty

B. Estimating Medium-Term Growth Expectations

14 INTERNATIONAL MONETARY FUND

Brazil: Consensus GDP Growth Forecasts

INTERNATIONAL MONETARY FUND 15

C. Estimation Results: Aggregate Model

Table 1. Brazil: Results—Aggregate Model

16 INTERNATIONAL MONETARY FUND

D. Robustness Checks: Firm-Level Model