Inflation Targeting in Brazil Lessons and Challenges November

Inflation Targeting in Brazil: Lessons and Challenges November 2002

The paper assesses the first three years of inflation targeting (IT) in Brazil (1999 -2002) • Overview of the three years • Challenges in the inflation targeting: - construction of credibility - change in relative prices - exchange rate volatility • Methodology to estimate effects of shocks • Institutional design of IT: core, escape clauses, tolerance intervals, targets 2

Main Results of Estimations • Inflation targets -> important coordinator of expectations • CB -> reacted strongly to inflation expectations • Degree of persistence of inflation -> reduction • Volatility inflation and output -> reduction 3

Main Results (cont. ) • Pass-through of Exchange Rate: - to IPCA: between 10% and 20% - to administered prices 2 times higher than to market prices - explains 38%-48% of inflation (2001 -Aug. 2002) 4

Overview of the three years

Macroeconomic Policy • Inflation Targeting • Floating Exchange Rate • Change in the fiscal regime primary surplus (%GDP): 0. 0% 1998 3. 7% 2001 3. 2% 1999 3. 5% Aug 2002 (12 -m) 3. 5% 2000 6

IT regime: successful, highly important for macroeconomic stabilization • 1999 and 2000: targets met • 2001 and 2002: targets not met - several shocks hitting the economy 12 10 8. 94 7. 67 8 5. 97 6 4 2 0 1999 2000 2001 2002 2003 2004 7

Volatilities • IT period: lower volatility of inflation, GDP, and interest rate Inflation Rate Period Real Plan Before Inflation Targeting 1994: 04 - 1999: 02 1996: 01 - 1999: 02 Output Average Standard Coefficient Average Standard (per year) Deviation of Variation (per year) Deviation Interest Rate Coefficient Average Standard of of (per year) Deviation Variation 10. 3 5. 8 9. 2 4. 8 0. 9 0. 8 2. 0 6. 3 5. 2 3. 2 2. 5 35. 4 28. 2 14. 1 6. 0 0. 40 0. 21 7. 1 3. 0 0. 4 2. 4 3. 5 18. 0 1. 4 0. 08 Inflation Targeting 1999: 03 - 2002: 02 8

Construction of Credibility

Reaction function of Central Bank • Interest rate reacts strongly to expected inflation • Monetary policy is conducted in a forward-looking basis and is tough against inflation 10

Estimation of Reaction Function of Central Bank Dependent Variable: Selic Interest Rate (Monthly Average) Coefficients and standard errors Regressors With Inflation Report With Market Inflation Expectations Constant Interest Rate (t-1) 17. 50*** (0. 36) 17. 56*** (0. 49) 0. 76*** (0. 07) 1. 04*** (0. 13) Interest Rate (t-2) Deviation of Expected Inflation Rate from Target 16. 48*** 16. 66*** (0. 64) (0. 70) 0. 72*** (0. 08) -0. 20** (0. 08) 1. 36*** (0. 18) -0. 56*** (0. 15) 1. 77** (0. 85) 1. 85 (1. 21) 1. 75** (0. 67) 1. 42* (0. 72) Output Gap (t-1) -0. 36*** (0. 09) -0. 38*** (0. 13) -0. 19* (0. 11) -0. 10 (0. 13) R-squared 0. 9287 0. 9415 0. 9270 0. 9538 Adjusted R-squared 0. 9220 0. 9340 0. 9186 0. 9464 LM Test for Autocorrelation of Residuals (p-values) 1 lag 0. 0352 0. 4877 0. 0056 0. 7313 4 lags 0. 2149 0. 6671 0. 0738 0. 5637 Notes: Standard error in parantheses. *, **, and *** indicate the coefficient is significant at 10%, 5%, and 1% level, respectively. 11

Estimation of Reaction Function of Central Bank Dependent Variable: Gap of Selic Interest Rate (Monthly Average) Coefficients and standard errors Regressors With Inflation Report With Market Inflation Expectations Constant -1. 51*** (0. 36) -1. 27*** (0. 36) -3. 29*** (0. 54) -3. 53*** (0. 65) Gap of Interest Rate (t-1) 0. 81*** (0. 06) 1. 08*** (0. 09) 0. 71*** (0. 08) 1. 34*** (0. 19) Gap of Interest Rate (t-2) -0. 25*** (0. 06) -0. 54*** (0. 15) Deviation of Expected Inflation Rate from Target 4. 98*** (0. 91) 4. 24*** (0. 77) 3. 71*** (0. 58) 3. 63*** (0. 68) Output Gap (t-1) -0. 30** (0. 12) -0. 34*** (0. 11) -0. 03 (0. 10) 0. 07 (0. 13) R-squared 0. 9654 0. 9768 0. 9694 0. 9798 Adjusted R-squared 0. 9622 0. 9738 0. 9658 0. 9765 LM Test for Autocorrelation of Residuals (p-values) 1 lag 0. 1309 0. 4205 0. 0079 0. 4261 4 lags 0. 4993 0. 0456 0. 4331 0. 0807 Notes: Standard error in parantheses. *, **, and *** indicate the coefficient is significant at 10%, 5%, an d 1% level, respectively. 12

Inflation Expectations: under control 12 -Month Ahead Expected Inflation and Inflation Target 9 8 7 5 4 3 2 1 Central Target Upper Limit Lower Limit Sep-02 Jul-02 May-02 Mar-02 Jan-02 Nov-01 Sep-01 Jul-01 May-01 Mar-01 Jan-01 Nov-00 Sep-00 Jul-00 May-00 Mar-00 0 Jan-00 % p. a. 6 Inflation Expectation 13

Reaction Function of Inflation Expectations • Inflation expectations respond to: - interest rate (positively); - inflation rate target (positively); - past inflation is not significant. • Inflation expectations are forward looking and the inflation targets have an important role 14

Estimation of Reaction Function of Inflation Expectations Dependent Variable: Market Inflation Rate Expectations (Adjusted) Coefficients and standard errors Regressors I Constant Market Inflation Rate Expectations (t-1) II IV -4. 23*** -4. 75*** -4. 22*** -4. 92*** (1. 30) (1. 31) (1. 32) (1. 35) 0. 24 (0. 18) Market Inflation Rate Expectations (t-2) 0. 48** (0. 19) 0. 27 (0. 18) -0. 39** (0. 15) 0. 50** (0. 19) -0. 40** (0. 15) Interest Rate (t-1) 0. 27*** (0. 07) 0. 29*** (0. 07) 0. 25*** (0. 08) 0. 27*** (0. 08) Inflation Rate Target (12 -Month Ahead) 0. 74*** (0. 22) 0. 96*** (0. 24) 0. 70*** (0. 23) 0. 95*** (0. 24) 0. 06 (0. 10) 0. 08 (0. 11) 12 -Month Inflation Rate (t-1) R-squared 0. 8978 0. 9030 0. 8993 0. 9053 Adjusted R-squared 0. 8855 0. 8861 0. 8826 0. 8838 LM Test for Autocorrelation of Residuals (p-values) 1 lag 0. 0663 0. 5196 0. 0652 0. 7730 4 lags 0. 0160 0. 1333 0. 0149 0. 1316 15

Change in the inflation dynamics • Simple Aggreggate Supply Curve - IT period: decrease in the degree of persistence in inflation 16

Estimation of Aggregate Supply Curve Dependent Variable: Monthly Inflation Rate Coefficients and standard errors Regressors Constant First Second Third Specification -0. 09* (0. 05) 0. 81** (0. 35) 0. 56*** (0. 15) 0. 81*** (0. 12) -0. 58*** (0. 21) -0. 12 (0. 12) -0. 30 (0. 20) -0. 10** (0. 05) R-squared Adjusted R-squared 0. 6431 0. 6269 0. 6766 0. 6538 0. 79** (0. 32) 0. 39*** (0. 14) 0. 61*** (0. 13) -0. 41** (0. 19) -0. 09 (0. 12) -0. 24 (0. 19) -0. 10** (0. 05) 0. 10*** (0. 04) 1. 08*** (0. 33) 0. 5537 0. 5055 LM Test for Autocorrelation of Residuals (p-values) 1 lag 4 lags 0. 1857 0. 0040 0. 8353 0. 1693 0. 5454 0. 1081 Dummy*Constant Inflation Rate (t-1) Dummy*Inflation Rate(t-1) 0. 74** (0. 36) 0. 38*** (0. 14) 0. 74*** (0. 09) -0. 59*** (0. 20) Inflation Rate (t-2) Dummy*Inflation Rate (t-2) Unemployment (t-1) Exchange Rate Change (t-1) (Twelve-Month Average) Dummy for 2000: 07 17

Change in relative prices

Relative increase in the “administered or monitored” prices • The group includes gasoline, cooking gas, electricity, telephone, urban bus • 30% of the consumer price index • Dynamics are different from the other prices: - international prices - greater pass-through from exchange rate - stronger backward-looking behavior 19

Jan-02 Jul-01 Jan-01 Jul-00 Jan-00 Jul-99 Jan-99 Jul-98 Jan-98 Jul-97 Jan-97 Jul-96 Jan-96 Jul-95 Jan-95 Jul-94 Jan-94 Jul-93 Jan-93 Jul-92 Jan-92 Ratio of Administered Prices to Market Prices 1. 6 1. 4 1. 2 1. 0 0. 8 20

Level of Prices – IPCA and selected items Dec-98=1 2. 8 2. 6 2. 4 2. 2 2. 0 1. 8 1. 6 1. 4 1. 2 IPCA Cooking Gas Electricity Gasoline Fixed Telephone Jun-02 Apr-02 Feb-02 Dec-01 Oct-01 Aug-01 Jun-01 Apr-01 Feb-01 Dec-00 Oct-00 Aug-00 Jun-00 Apr-00 Feb-00 Dec-99 Oct-99 Aug-99 Jun-99 Apr-99 Feb-99 Dec-98 1. 0 Urban Bus 21

Granger Causality Test: Relative Prices and IPCA Sample: 1994: 12 – 2002: 06 Null Hypothesis 1 lag 3 lags F-statistic P-value Relative Prices do not Granger Cause Inflation Rate 2. 89 0. 0926 3. 15 0. 0293 Inflation Rate does not Granger Cause Relative Prices 16. 71 0. 0001 5. 35 0. 0020 22

Exchange Rate Volatility 1999: 07 – 2002: 06 Average Monthly Increase 1. 2% p. m. s. e. : 3. 6

Jul-02 May-02 Mar-02 Jan-02 Nov-01 Sep-01 Jul-01 May-01 Mar-01 Jan-01 Nov-00 Sep-00 Jul-00 May-00 Mar-00 Jan-00 Nov-99 Sep-99 Jul-99 May-99 Mar-99 Jan-99 Nov-98 Sep-98 Jul-98 May-98 Mar-98 Jan-98 Exchange Rate Level (R$/US$) 3. 50 3. 00 2. 50 2. 00 1. 50 1. 00 24

Exchange Rate Depreciation and Inflation in 2001 and Jan. -Aug. /2002 30 24. 1 25 20. 9 20 15 10. 4 10 7. 6 7. 7 6. 5 4. 8 5 3. 7 0 Exchange Rate Depreciation Administered Prices Inflation 2001 Inflation Rate (IPCA) Market Prices Inflation 2002 (up to August) 25

Contributions to Inflation in 2001 Percentage of the Total and Percentage Variation in the Period Inertia 10% 0. 7 2. 4 Exchange Rate Pass. Through 38% Market Prices Inflation Excluding Exchange Rate Pass. Through and Inertia 28% 2. 9 1. 7 Administered Prices Inflation Excluding Exchange Rate Pass. Through and Inertia 24% 26

Contributions to Inflation in Jan-Aug/2002 Percentage of the Total and Percentage Variation in the Period Inertia 17% 0. 8 Market Prices Inflation Excluding Exchange Rate Pass-Through and Inertia 22% 1. 1 0. 6 Exchange Rate Pass. Through 48% 2. 3 Administered Prices Inflation Excluding Exchange Rate Pass. Through and Inertia 13% 27

VAR estimation: Impulse Response Functions • Output, adm. prices, market prices, EMBI+, exchange rate, interest rate Response of ADMP to ER 1. 2 Response of FREEP to ER 6 0. 5 1. 0 5 0. 4 0. 8 4 3 0. 6 2 0. 4 1 0 0. 1 0. 2 -1 0. 0 -0. 2 1 2 3 4 5 6 7 8 9 10 11 12 Response of ER to ER -2 -3 -0. 1 1 2 3 4 5 6 7 8 9 10 11 12 28

Pass-Through from Exchange Rate Model Estimation Administered Prices Market Prices Consumer Price Index (IPCA) Structural Model and Priori Information 25 12 16 VAR Estimation 20 8 14 12 5 8 Simple AS Curve 21 9 (IT period) 29

Methodology to calculate effects of shocks to administered prices

Strategy of Central Bank • Central Bank accomodates primary effects of shocks to administered prices but not secondround effects • To calculate the primary effect of the shocks to administered prices, the effects of the exchangerate change and inflation inertia are excluded: 31

Strategy of Central Bank • and only part of inertia is neutralized in the current year • Therefore, the components concerning the primary effect of shocks to administered prices and part of the inflation inertia are excluded from the expected inflation. The new value is then compared to the inflation target. 32

Example of Calculation of Inflation Excluding Inertia Effect from Previous Year and the Primary Effect of the Shock to Administered Prices Item % (a) Inflation forecast for year t 5. 0 (b) Contribution of year t-1 inertia to market price inflation in year t 0. 6 (c) Inertia from year t-1 to be neutralized in year t +1 (= b/2) 0. 3 (d) Administered price inflation forecast for year t 9. 0 (e) Contribution of administered price inflation above the target (= (d - p target )wadm) 1. 2 (f) Inertia effects of year t-1 on year t administered price inflation 0. 2 (g) Exchange rate impact on administered price inflation 0. 5 (h) Primary effect of the shock to administered prices (= e - f -g) 0. 5 (i) Inflation forecast excluding the inertia effect and the primary effect of the shocks to administered prices (= a - c - h) 4. 2 33

Institutional Design of Inflation Targeting • • Core Escape clauses Tolerance intervals Establishment of targets

Headline versus core measure • Core by exclusion – two problems: - agents are concerned about the whole basket of consumption - In Brazil, in the past, there were some exclusions in the official price index that would affect the credibility if a core by exclusion were adopted • Only Canada and Thailand adopt it. Australia, New Zealand, and Czech Republic have abandoned it (see Ferreira and Petrassi, 2002) 35

Inflation Rate Jul-02 Apr-02 Jan-02 Oct-01 Jul-01 Apr-01 Jan-01 Oct-00 Jul-00 Apr-00 Jan-00 Oct-99 Jul-99 Apr-99 Jan-99 Oct-98 Jul-98 Apr-98 Jan-98 Oct-97 Jul-97 Apr-97 Jan-97 Oct-96 Jul-96 Apr-96 Jan-96 Inflation Rate and (trimmed mean) Core Inflation Rate and Core Inflation 1996: 01 - 2002: 08 (% p. m. ) 2. 0 1. 5 1. 0 0. 5 0. 0 -0. 5 Core Inflation 36

Trimmed mean core: good predictor of trend inflation Granger Causality Test: Core Inflation and Inflation Rate (IPCA) Sample: 1996: 01 - 2002: 06 Null Hypothesis 1 lag 2 lags F-statistic P-value Core Inflation does not Granger Cause Inflation Rate 4. 83 0. 0311 7. 14 0. 0015 Inflation Rate does not Granger Cause Core Inflation 0. 04 0. 8479 0. 98 0. 3797 37

Escape clauses • the circumstances under which central banks can justify non-fulfillment of targets are set in advance (natural disasters, change in international prices, in indirect taxes, etc. ) – New Zealand, South Africa, Czech Republic, Switzerland • the adoption of escape clauses could signal central bank would be lenient in the context of a credibility under construction 38

Tolerance intervals (bands) • Reasons: - limits of forecasting models - unexpected shocks - lags in the effects of monetary policy • Brazil: short period of stabilization, frequent and large shocks - tolerance intervals: enlarged from 2 p. p. to 2. 5 p. p. above and below the central target 39

Establishment of targets • Brazil: target is set in June of year t for t+2 • Change in the target for 2003: from 3. 25% to 4. 0% • Loss of credibility of change versus loss of credibility of pursuing a unlikely target 40

Main conclusions • The inflation-targeting regime in Brazil is relatively new, but has shown to be important to achieve low levels of inflation rate even in a context of large shocks • Regime has faced many challenges: 1. construction of credibility - CB has reacted strongly to inflation expectations - inflation expectations under control - reduction in the degree of inflation persistence 41

Main conclusions 2. change in relative prices -> inflation - CB has developed methodology to estimate primary - effects of administered prices 3. exchange rate volatility -> inflation - Institutional design - CB has enlarged the band size 42